Impact of CEO Political Ideology on Labor Cost Reductions and Payout Decisions During the COVID-19 Pandemic
Summary
An open-access article by Ali Bayat, Marc Goergen, Panagiotis Koutroumpis and Xingjie Wei in the Journal of Corporate Finance 90 (2025) 102692 on whether CEO political ideology affected how S&P 500 firms reacted to the COVID-19 pandemic in 2020. Using a hand-collected dataset, the authors hypothesize that conservative CEOs were more likely to aggressively reduce labor costs while still meeting dividend expectations, and they report that their findings support this hypothesis. They also find that conservative CEOs used temporary downsizing to avoid earnings losses, enabling them to meet dividend expectations. The article was received 8 April 2024 and accepted 25 October 2024, is published under a Creative Commons Attribution 4.0 (CC-BY) license, and ends with its reference list.
Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used
- Issuer
- Source documents
- Document type
- The Impact Of Ceo Political Ideology On Labor Cost Reductions And Payout Decisions During
- Case
- The Impact Of Ceo Political Ideology On Labor Cost Reductions And Payout Decisions During The Covid 19 Pandemic
Cite this document
“Impact of CEO Political Ideology on Labor Cost Reductions and Payout Decisions During the COVID-19 Pandemic.”
Original at doi.org. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/the-impact-of-ceo-political-ideology-on-labor-cost-reductions-and-payout-decisions-during-the-covid-19-pandemic/
, SHA-256 02db3dcad58ad748e83bec0d55bd4c1feed001857d68772472300ccc9e9fffa1.Full text
The impact of CEO political ideology on
labor cost reductions and payout
decisions during the COVID-19 pandemic
Article
Published Version
Creative Commons: Attribution 4.0 (CC-BY)
Open Access
Bayat, A., Goergen, M. ORCID: https://orcid.org/0000-0003-
4391-2651, Koutroumpis, P. ORCID: https://orcid.org/0000-
0002-2281-7236 and Wei, X. ORCID: https://orcid.org/0000-
0002-6064-7290 (2025) The impact of CEO political ideology
on labor cost reductions and payout decisions during the
COVID-19 pandemic. Journal of Corporate Finance, 90.
102692. ISSN 1872-6313 doi: 10.1016/j.jcorpfin.2024.102692
Available at https://centaur.reading.ac.uk/124811/
It is advisable to refer to the publisher’s version if you intend to cite from the
work. See Guidance on citing .
To link to this article DOI: http://dx.doi.org/10.1016/j.jcorpfin.2024.102692
Publisher: Elsevier
All outputs in CentAUR are protected by Intellectual Property Rights law,
including copyright law. Copyright and IPR is retained by the creators or other
copyright holders. Terms and conditions for use of this material are defined in
the End User Agreement .
www.reading.ac.uk/centaur
CentAUR
Central Archive at the University of Reading
Reading’s research outputs online
Journal of Corporate Finance 90 (2025) 102692
Contents lists available at ScienceDirect
Journal of Corporate Finance
journal homepage: www.elsevier.com/locate/jcorpfin
The impact of CEO political ideology on labor cost reductions and
payout decisions during the COVID-19 pandemic
Ali Bayat a , Marc Goergen b,c ,∗, Panagiotis Koutroumpis d , Xingjie Wei e
a
University of Aberdeen, UK
b
IE Business School, IE University, Spain
c
European Corporate Governance Institute, Belgium
d
University of Sussex Business School, University of Sussex, UK
e
Centre for Decision Research, Leeds University Business School, University of Leeds, UK
ARTICLE INFO ABSTRACT
Editor: T Wang Using a hand-collected dataset, we study whether CEO political ideology affected S&P 500
firms’ reactions to the COVID-19 pandemic in 2020. During the pandemic, CEOs had the
JEL classification:
option to distribute the pain of the pandemic’s impact onto shareholders by paying lower
G35
G34
dividends, onto the workforce by reducing labor costs, or to share the pain. We hypothesize
M51 that conservative CEOs were more likely to aggressively reduce labor costs while still meeting
dividend expectations. Conversely, other CEOs would have been less likely to meet dividend
Keywords: expectations and less likely to reduce labor costs. Our findings support this hypothesis. We also
CEO political ideology
find that during the pandemic, conservative CEOs used temporary downsizing to avoid earnings
Dividend policy
losses, enabling them to meet dividend expectations.
Downsizing
Labor cost reductions
Stakeholder management
COVID-19 pandemic
As the pandemic squeezes big companies, executives are making decisions about who will bear the brunt of the sacrifices,
and in at least some cases, workers have been the first to lose, even as shareholders continue to collect.
[Peter Whoriskey, The Washington Post, May 5, 2020]
1. Introduction
In the face of extreme and sudden economic adversities, how do CEOs reconcile the priorities of their shareholders with
those of their employees? Scholarly work to date has predominantly focused on the manner in which the personal ideologies and
preferences of CEOs influence their decisions regarding shareholders and employees in periods of economic stability (e.g., Di Giuli
and Kostovetsky, 2014; Chin et al., 2013; Adams et al., 2011). Three notable and recent examples are Bizjak et al. (2022), Keum
and Meier (2023), and Guenzel et al. (2023). The former examines how CEOs with different political leanings balanced the interests
of shareholders and public health during the pandemic. The latter two explore the impact of CEO social preferences on layoff
decisions. However, their analyses stop short of the 2008 crisis and the COVID-19 pandemic, respectively. Consequently, the role
∗ Corresponding author at: IE Business School, IE University, Spain.
E-mail addresses: ali.Bayat@abdn.ac.uk (A. Bayat), marc.goergen@ie.edu (M. Goergen), p.koutroumpis@sussex.ac.uk (P. Koutroumpis), x.wei1@leeds.ac.uk
(X. Wei).
https://doi.org/10.1016/j.jcorpfin.2024.102692
Received 8 April 2024; Received in revised form 23 October 2024; Accepted 25 October 2024
Available online 16 November 2024
0929-1199/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
of CEO decision-making under sudden and extreme market conditions, where effective leadership is crucial in balancing the needs
of the employees and shareholders, is not well documented.1
We attempt to fill the identified gap in the literature using a unique and granular dataset related to how S&P 500 firms adjusted
labor costs during the 2020 COVID-19 pandemic. The period marked by the pandemic precipitated a substantial decline in earnings
for a considerable proportion of firms. In contrast to the 2008 financial crisis, which was preceded by warning signs (e.g., Pettifor,
2006), not only was the pandemic unexpected, but the contraction of the global economy it caused was also much more sudden and
severe (IMF, 2020). Hence, the pandemic forced CEOs of affected firms to take rapid and potentially extreme measures to reduce
costs and preserve liquidity, which may have been primarily achieved by reducing labor costs, cutting dividends, or both.
We examine the influence of CEO political ideology on how the pain caused by the pandemic was distributed across employees
and shareholders. During normal times, liberal-leaning CEOs prioritize the interests of nonfinancial stakeholders more than their
conservative counterparts do; this is well documented in the literature.2 However, it remains unclear whether such behavior is
sustained during economic shocks.
Building on this foundation, we investigate the role of CEO political preferences during the pandemic. Our investigation centers
on whether the financial pain was borne primarily by the employees, through either temporary or permanent cuts in labor costs;
primarily by shareholders, through reduced dividends; or equally by employees and shareholders. We posit that CEO political
ideology shapes CEOs’ prioritization of stakeholder interests, impacting decisions on labor cost adjustments and dividend payments.
We conjecture that conservative-leaning CEOs are more inclined to protect shareholder interests, potentially at the expense of the
workforce, by favoring strategies that may include permanent downsizing. In contrast, CEOs with less conservative ideologies may
opt for temporary reductions in labor costs, with the aim of preserving employment.
We argue that CEOs may have used dividend cuts as an alternative way of dealing with liquidity crises that they may have
faced during the pandemic. This may be the case for the following two reasons. First, given the sudden and exogenous nature of the
pandemic, it would have been unlikely for investors to blame CEOs for dividends falling short of expectations. Thus, cutting dividends
during the pandemic would have been a legitimate and possible option. Dividend cuts may even have been the preferred option
compared to the alternative of reducing labor costs via, for example, laying off employees for reasons beyond employee performance.
In contrast, under normal circumstances, CEOs are dissuaded from cutting dividends because hefty penalties are typically meted out
by capital markets for such cuts.3
Second, dividend cuts could have been considered a plausible alternative to options such as external financing or reductions in
capital expenditures. Indeed, when internal financing is insufficient to cover investments, firms can (i) raise external capital, (ii)
reduce investment while keeping their payout policy unchanged, or (iii) reduce the payout while increasing investment (Iyer et al.,
2016). Given that the cost of external financing is prohibitive during a crisis, the first alternative option can be easily discarded for
the COVID-19 pandemic. Hence, only the two latter options remain. We argue that the perceived temporary nature of the COVID-19
pandemic would have increased CEOs’ propensity to reduce dividends compared to reducing capital expenditures, as the latter type
of reduction would have turned the otherwise temporary effect of the pandemic into a longer-term effect. In addition, dividend
changes can be implemented relatively quickly, and they therefore have a relatively rapid impact on a firm’s liquidity, which
is not the case for reductions in capital expenditures. In support of our argument, Campello et al. (2010), who surveyed 1,050
chief financial officers (CFOs) during the 2008 financial crisis, report that the respondents expected the smallest cuts to capital
expenditures. Further support is provided by Iyer et al. (2016), who find no link between cash dividends and capital expenditures
under normal circumstances. The lack of a negative correlation suggests that firms are not constrained from cutting dividends and
may opt to preserve capital expenditures instead.
Past research has typically measured reductions in labor costs through a drop in the number of employees using firm-level
employment counts.4 In contrast, this study benefits from a hand-collected and highly granular dataset related to how firms reduced
their labor costs during the pandemic. The dataset was collected from filings of Form 8-K, which is the form that firms must file with
the Securities and Exchange Commission (SEC) to disclose material events (in accordance with Section 409 of the Sarbanes-Oxley
Act of 2002). Hence, our data not only tell us whether a firm reduced its labor costs but also how this reduction was performed
(e.g., labor cost reductions via temporary or permanent downsizing). Importantly, our data also include information about reductions
in labor costs that did not result in reductions in employee numbers (e.g., reductions in wages and reductions in the number of hours
worked). This granularity enables a precise examination of CEO actions during the pandemic; therefore, it does not suffer from the
limitations of indirect and aggregate measures of labor cost reductions.
We find evidence that conservative CEOs were more likely to meet – or even exceed – dividend expectations, while also being
more likely to reduce labor costs. In other words, conservative CEOs were more likely to choose actions that resulted in employees
bearing the pain from the economic shock caused by the pandemic while meeting investors’ dividend expectations from the end of
2019, the year preceding the pandemic. In contrast, the remaining CEOs – and not just the liberal ones – were more likely to opt
1 There is a body of literature on how CEO political ideology affected firms around the 2008 financial crisis (see, e.g., Campbell et al., 2022, for a recent
overview). However, this literature tends to focus on firms’ and banks’ risk-taking behavior or financial performance before and during the crisis.
2 See Chin et al. (2021) for a recent review.
3 Such penalties include a drop in the share price (see, e.g., Healy and Palepu (1988), Michaely et al. (1995), Benartzi et al. (1997), Jensen et al. (2010)) and
institutional investors voting with their feet by reducing their stock holdings (see, e.g., Parrino et al. (2003)). Dividend cuts have also been shown to increase the
likelihood of the CEO being dismissed (Parrino et al., 2003; Schaeck et al., 2011) and to lower the CEO’s future number of seats on other firms’ boards (Kaplan
and Reishus, 1990).
4 See Guenzel et al. (2023) for a discussion.
2
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
for reactions to the pandemic that either shared the pain between shareholders and employees or made shareholders bear all the
pain. We also find some evidence that conservative CEOs were more likely to use temporary reductions in labor to avoid negative
earnings per share, which then enabled them to meet dividend expectations. Finally, we do not find any evidence that conservative
CEOs reduced labor costs to maintain their stock repurchase programs during or after the pandemic. Our results are upheld via a
battery of robustness tests.
Our paper contributes to our understanding of CEO behavior during periods of crisis. It contributes to an emerging literature on
how CEOs behave during extreme states of the world (e.g., Matsa and Miller, 2014; Yonker, 2017; Cejnek et al., 2021; Afzali et al.,
2022), as well as to our knowledge of the importance of the political preferences of CEOs in decision-making (e.g., Hutton et al.,
2014; Elnahas and Kim, 2017; Unsal et al., 2016; Francis et al., 2016; Gupta et al., 2018). This is achieved by documenting that
CEO political ideology is a key determinant of how the financial burden caused by the pandemic is allocated across shareholders
and employees, confirming that the effect of political ideology extends to times of crisis. More importantly, in this vein, our paper
illustrates that the dividend decision and the decision to reduce labor costs in response to a crisis are closely linked to each other;
therefore, they should not be studied in isolation, as is normally done in extant literature.
Closest to our study is the research by Afzali et al. (2022); these authors examine the determinants of a CEO’s decision to share
the pain of the pandemic with employees. They find that performance prior to the pandemic and the firm’s exposure to the pandemic
are the main determinants of the CEO sharing the pain with the employees. In contrast, our study focuses on how CEOs apportion
the pain from the pandemic across shareholders and employees. Note that contrary to our study, Afzali et al. (2022) limit themselves
to furloughing and reducing the number of workers. Again, the present study also takes into account other ways of reducing labor
costs, which did not affect the number of workers.
The remainder of this paper is organized as follows. The next section provides a brief summary of the events surrounding the
COVID-19 pandemic, including the economic shortfall it caused and the U.S. government’s reaction to it. Following this, Section 3
describes the sample selection and methodology. Section 4 then focuses on the empirical analysis, including the robustness tests
and further analysis. Finally, Section 5 concludes.
2. The COVID-19 pandemic
The first identified COVID-19 case – that is, the first case of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) –
can be traced back to December 2019 and was reported in China5 ; the first American case was reported soon thereafter, on January
20, 2020. On January 31, the then health and human services secretary Alex M. Azar declared a public health emergency.6 The
pandemic hit the U.S. particularly hard; it was the country that suffered the highest number of COVID-19 cases and related deaths.7
The economic fallout from the pandemic was not only rapid but also substantial (Bartik et al., 2020), as evidenced by a drop in the
S&P 500 index from 3225 on January 31 to 2237 on March 23. The index then gradually recovered to reach 3756 on December
31.8 This was combined with an increase in the unemployment rate from 3.5% in January and February to 14.8% in April, a greater
peak than that reached during the Great Recession caused by the 2008 financial crisis (BLS, 2012). In July 2020, the unemployment
rate somewhat recovered, dropping to 10.2%. At the end of calendar year, it was down to 6.7%.9 To sum up, the economic fallout
was mainly experienced during the first two quarters of 2020. The effects of the pandemic varied across industries, with the retail,
leisure, hospitality, and travel industries suffering the most (Chetty et al., 2022; Bartik et al., 2020).10
The U.S. government’s main response to the pandemic was the Coronavirus Aid, Relief, and Economic Security (CARES) Act,
which came into effect on March 27, 2020. The CARES Act introduced a number of measures targeting not only individuals and
small businesses but also large corporations to help them deal with the economic consequences of the COVID-19 pandemic. More
specifically, the Act focused on relief measures for healthcare providers, manufacturers, and distributors, such as loans, tax credits,
tax deductions, and tax deferrals. Such measures included steps to make it easier for corporations to avoid downsizing, such as
the deferral of the employer’s share of Social Security taxes for up to two years and a refundable employee retention tax credit.
A condition of receiving support under the CARES Act was that firms were not permitted to conduct involuntary terminations or
furloughs. For example, the Payroll Support Program to Air Carriers and Contractors, published in March 2020, required applicant
firms to refrain from conducting involuntary layoffs or furloughs for six months.11
As Congdon and Vroman (2022) document, there were 30 million unemployment insurance (UI) claims during the pandemic;
during the Great Recession, the equivalent number did not exceed 10 million. In addition, the Pandemic Unemployment Assistance
(PUA) program provided unemployment benefits to millions of workers who would not normally have been eligible for UI (e.g., gig
workers and workers with caring responsibilities). Government support was also more generous in that additional benefits were
5 See WHO (2021).
6 See https://docs.house.gov/meetings/IF/IF14/20200226/110589/HHRG-116-IF14-20200226-SD002.pdf, accessed on January 30, 2023.
7 Johns Hopkins Whiting School of Engineering — Center for Systems Science and Engineering, JHU CSSE COVID-19 Project, https://systems.jhu.edu/research/
public-health/ncov/, accessed on January 30, 2023.
8 See https://fred.stlouisfed.org/series/SP500, accessed on September 29, 2022.
9 See https://fred.stlouisfed.org/series/UNRATE, accessed on September 29, 2022.
10 See Chetty et al. (2022) for a detailed study on the effects of the pandemic on consumer spending and revenues and employment for small businesses. This
study finds that consumer spending dropped the most in the In-Person Services industry sector, with the Hotels & Food and Transportation industry subsectors
being hit particularly hard. See also Bartik et al. (2020), Alekseev et al. (2023).
11 See https://home.treasury.gov/system/files/136/Payroll-Support-Procedures-Form-FINAL.pdf, accessed on January 27, 2023.
3
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
provided by the Federal Pandemic Unemployment Compensation (FPUC) program, amounting to $600 for the first week and then
$300 for following weeks.
Finally, research suggests that the political leanings of citizens in general and corporate decision makers may have affected
how they reacted to the pandemic. Although Vorsatz (2022) found no difference between Republican and Democrat mutual fund
teams – both were more risk averse during the COVID-19 pandemic when compared to nonpartisan fund teams – Cookson et al.
(2020) and Benton et al. (2022) found that conservative investors and CEOs were more optimistic about the pandemic. Rothwell
and Makridis (2020) and Canes-Wrone et al. (2020) provide supporting evidence for this, showing that citizens of red states were
less worried about COVID-19, practiced less social distancing and social isolation, and were less likely to wear a face mask. They
were also more likely to visit their workplace and retail stores.
3. Sample selection, data collection, and methodology
3.1. Sample selection and data collection
Given the focus of this paper on investigating whether CEO political ideology explains whether firms cut their dividends, reduced
labor costs – including the downsizing of the workforce – or did both in response to the pandemic, it is important that the firms
in question paid a dividend before the 2020 pandemic. Therefore, we focus on S&P 500 firms because they are more likely to be
dividend payers.12 We obtained the list of the S&P 500 firms from Bloomberg for the year 2020. After excluding the 19 companies
with headquarters located outside the U.S., the initial sample consists of 481 S&P 500 firms.
For the initial sample, we collect CEO data from Execucomp and BoardEx, then match these data with firm financial and
accounting data obtained from Compustat. We then complement the data with political donations data obtained from the Federal
Election Commission (FEC) and labor cost reduction data obtained from Form 8-K filings submitted by firms during the four quarters
of calendar year 2020. After merging the financial and accounting data with the political donation and labor cost reduction data,
as well as discarding firms with missing observations, we obtain a final sample of 459 firms for use in the regression analysis.13
We measure CEO political ideology in terms of political donations. We follow previous work (see, e.g., Hutton et al., 2014) by
measuring CEO political ideology via the political donations made by the CEOs during their lifetime up to and including calendar
year 2020.14 We use a matching algorithm, combined with a manual check, to filter out CEOs from other donors with similar names,
using information about donor occupation, employer, and address. For each year, we aggregate the donations to obtain the dollar
value of the total contributions to each party made by each CEO. We only consider CEOs’ direct contributions to the Republican
and Democratic parties, and we exclude indirect donations made via a political action committee (PAC), as CEOs themselves are
not fully in control of the choice of recipients for the donations made by a PAC (Hutton et al., 2014). Hence, the donations made
by a PAC are likely to be a reflection of the political ideology of the firm’s workforce and not necessarily a true reflection of its
CEO’s political leaning.
Quarterly financial and accounting data are sourced from Compustat. We collected quarterly rather than annual data because
the former provide us with more granularity. Indeed, a dividend falling below the expected levels in one quarter of 2020 could have
been reversed in a subsequent quarter. While such temporary changes are reflected in the quarterly data, this may not necessarily
be the case with the annual data.
The data on labor cost reductions were obtained from Form 8-K filings made by S&P 500 firms during the four quarters of
calendar year 2020 (i.e., 7577 filings in total). In accordance with Section 409 of the Sarbanes-Oxley Act of 2002, Form 8-K is the
form that firms must file with the SEC to disclose events that are likely to be material. Using Form 8-K filings enables us to collect
granular data, including data on the way firms reduced their labor costs. While some of these reductions in labor costs would have
resulted in downsizing – that is, reductions in the number of employees – our data enable us to assess whether the downsizing
was temporary or permanent; this cannot be deduced by collecting data on annual employment figures. Importantly, our data on
downsizing also include reductions in labor, such as reductions in the weekly working hours, which do not affect the number of
employees.
The data on labor cost reductions was collected via a four-step process. First, we manually checked a subset of Form 8-K filings
to generate a list of ‘‘core’’ keywords related to labor cost reductions. Second, we added semantically similar words to the list of
keywords using a word embedding model (Pennington et al., 2014) to include as many words as possible related to labor cost
reductions. We followed a similar process to identify words related to COVID-19 and to seeking assistance under the CARES Act.
Third, we conducted a fuzzy search15 to identify Form 8-K filings containing at least one of the keywords related to COVID-19, as
well as a number of keywords possibly related to labor cost reductions and applications made under the CARES Act. These ‘‘core’’
keywords are listed below. It should be noted that our text search includes not only these words but also a wider list with word
variations and semantically similar words obtained via the word embedding model (Pennington et al., 2014).
12 Out of the 481 U.S.-headquartered S&P 500 firms, only 75 did not pay a dividend in the pre-pandemic year 2019, while 64 of these 75 firms did not pay
a dividend in 2020 either. Fifty-five percent of the firms that did not pay a dividend in at least one of these two years were in the high-tech industry.
13 Further details about the sample selection process can be found in Appendix A.
14 Please note that there were only 112 donations (out of a total 12,246 donations) made by CEOs in calendar year 2020; excluding these donations does not
materially change our political ideology measures.
15 A fuzzy search is a text mining technique of approximate string matching that may be less than 100% perfect when finding correspondences between
keywords and words in a paragraph of text (Levenshtein, 1966).
4
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
• Keywords for COVID-19: COVID-19, coronavirus, pandemic, epidemic, and health crisis.
• Keywords for labor cost reductions: workforce, headcount, staff, employees, personnel, labor force, response, downsizing, furlough,
shutdown, close, layoff, restructure, retrench, redundancy, compulsory, voluntary, temporary, shorten, early leave, forgo, and reduce.
• Keywords for CARES: coronavirus aid, CARES act, relief, assistance, payroll support, loan, treasury department, and economic
security.16
Finally, we manually checked all the identified Form 8-K filings for the presence of the above keywords. To ensure data
consistency, we also cross-checked the information provided in these filings on applications made under the CARES Act with
information from the website of the U.S. Treasury Department.17
3.2. Methodology
3.2.1. Regression models and key variables
To test the validity of our hypothesis that CEO political ideology would affect the way a CEO reacted to the 2020 pandemic, we
estimate multinomial logit regressions. While full details about the multinomial logits can be found in Table C.1 in Appendix C, in
a nutshell, the various multinomial logits enable us to estimate the likelihood of a (more) conservative CEO choosing a (more)
shareholder-friendly reaction to the pandemic as opposed to a (more) employee-friendly reaction. We consider reactions that
prioritize safeguarding the dividend – more precisely, those meeting dividend expectations – over maintaining employee numbers
as shareholder-friendly, while reactions that avoid reductions in labor costs are considered employee-friendly.
The question arises as to which comparator should be used for the quarterly dividends in calendar year 2020. On the one hand,
it can be argued that the comparator should be the dividend per share for the equivalent quarter from the previous year (year 2019),
as suggested by Lintner (1956). For example, the dividend for the second quarter of 2020 should then be compared to the dividend
for the second quarter of 2019. If the former is lower than the latter, this would constitute a dividend cut. On the other hand, it can
be argued that the counterfactual for the pandemic is not the year 2019; preceding the pandemic, investors may have expected the
dividend to increase in 2020 in line with earnings per share that were also expected to increase. Hence, we opted to compare each
quarterly dividend per share (DPS) in 2020 to the expected DPS for the respective quarter. We use the difference between the actual
DPS in a given quarter of 2020 and the arithmetic mean of analysts’ forecasts for the DPS in 2019. These data are obtained from
I/B/E/S; the forecasts we use are the latest made during the calendar year 2019. It should be noted that using the latest dividend
forecasts made in 2019 could cause a bias as, for example, the forecast for the first quarter of 2020 may be more accurate than the
forecast for the fourth quarter of 2020. Nevertheless, the economic fallout from COVID-19 was mainly experienced during the first
quarter, with the U.S. economy recovering steadily in quarters 3 and 4 of calendar year 2020. Therefore, we would expect this bias
to be lowest during the first (two) quarter(s), when firms were more likely to adjust their dividends and labor costs.
While the actual change in the dividend in quarter q of year t, 𝑞 𝑡, is equal to:
𝐷𝐼 𝑉𝑞 𝑡 − 𝐷𝐼 𝑉𝑞 ,𝑡−1 ,
we compare the actual dividend to the expected dividend, that is, the dividend forecast made in t-1, as follows:
𝐷𝐼 𝑉𝑞 𝑡 − 𝐸𝑡−1 [𝐷𝐼 𝑉𝑞 𝑡 ].
It should be noted that our approach is equivalent to comparing the actual change in the dividend to the expected change in the
dividend, that is:
𝐷𝐼 𝑉𝑞 𝑡 − 𝐷𝐼 𝑉𝑞 ,𝑡−1 − {𝐸𝑡−1 [𝐷𝐼 𝑉𝑞 𝑡 ] − 𝐷𝐼 𝑉𝑞 ,𝑡−1 }.
While the above discussion focused on dividends, we also account for stock repurchases in the following analysis. The question
arises as to whether we should also take into account special dividends. We follow DeAngelo et al. (2000) to identify firms with
special dividends in the Center for Research in Security Prices (CRSP) database. We classify a dividend as a special dividend if it
has a distribution code of 1262 or 1272. As we only identify 7 firms with special dividends in 2019, 12 firms in 2020, and 13 firms
in 2021, it is not possible to perform a regression analysis. More importantly, these numbers suggest that firms that reduced their
labor costs were unlikely to use special dividends to disperse these cost savings to their shareholders.
Again, to test the validity of our main hypothesis – that conservative CEOs were more likely to let their employees bear the brunt
of the economic fallout from the pandemic – we distinguish between entirely shareholder-friendly reactions, entirely employee-
friendly reactions, and reactions that favor neither. In the following, we refer to these reactions as ‘‘no pain’’, ‘‘shareholder pain’’,
‘‘no employee pain and shareholder joy’’, ‘‘employee pain’’, and ‘‘shared pain’’, as defined below:
• ‘‘No pain’’: No reduction in labor costs and paying out the expected dividend;
• ‘‘Shareholder pain’’: No reduction in labor costs and paying out a dividend below the expected dividend, which implies that
the shareholders bear all the pain;
16 For each keyword, the search also uses variations of the keyword, such as ‘reducing’, ‘reduction’, and ‘reduced’ for ‘reduce’.
17 https://home.treasury.gov/policy-issues/cares, accessed on January 27, 2023.
5
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
• ‘‘No employee pain and shareholder joy’’: No reduction in labor costs and paying out a dividend above the expected dividend
and/or repurchasing shares;18
• ‘‘Employee pain’’: Reducing labor costs while maintaining the dividend, i.e., meeting the expected dividend, which implies
that the employees bear all the pain; and
• ‘‘Shared pain’’: Reducing labor costs and dividends below the expected dividend.19
To estimate the likelihood of the reactions to the pandemic depending on CEO political ideology, we estimate multinomial logits
based on Eq. (1):
𝑃 𝑟𝑜𝑏(𝑅𝑒𝑎𝑐 𝑡𝑖𝑜𝑛 𝑡𝑦𝑝𝑒 𝑡𝑜 𝑝𝑎𝑛𝑑 𝑒𝑚𝑖𝑐) = 𝛼 + 𝛽1 ∗ 𝐶 𝐸 𝑂 𝑝𝑜𝑙𝑖𝑡𝑖𝑐 𝑎𝑙 𝑖𝑑 𝑒𝑜𝑙𝑜𝑔 𝑦 + 𝛽2 ∗ 𝐿𝑜𝑠𝑠
+ 𝛽3 ∗ 𝐿𝑜𝑠𝑠 ∗ 𝐶 𝐸 𝑂 𝑝𝑜𝑙𝑖𝑡𝑖𝑐 𝑎𝑙 𝑖𝑑 𝑒𝑜𝑙𝑜𝑔 𝑦 (1)
+ 𝛽4 ∗ 𝐶 𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒𝑠 + 𝜖
The dependent variable of the multinomial logits is explained in detail in Table C.1 in Appendix C.
We run three different multinomial logits because the above pairs of alternatives require three different base cases (i.e., ‘‘share-
holder pain’’, ‘‘shared pain’’, and ‘‘no pain’’). Unless otherwise stated, all variables are based on calendar year 2020. Although the
quarterly data for 2020 are aggregated, our empirical analysis still benefits from the granularity of the quarterly data. Indeed, the
DPS falling below expectations in the 2020 calendar year is defined as a DPS in any of the four quarters of 2020 that is below the
expected DPS for the equivalent quarter. In turn, a dividend exceeding expectations in 2020 is defined as a DPS for any of the four
quarters of 2020 that exceeds the expected DPS for the equivalent quarter. Firms with dividends both falling below expectations in
some quarters and dividends exceeding expectations in the remaining quarters of 2020 are classed as firms whose dividend fell below
expectations in 2020. Given the context of the pandemic, this approach takes into account that firms may not have met dividend
expectations during the first two or three quarters of 2020 but then met dividend expectations from 2019 once their earnings had
recovered sufficiently.20
It should be noted that the dependent variable in Eq. (1) takes into account whether the firm reduced its labor costs. However,
it does not distinguish between the various ways of doing so. When it comes to downsizing, this could take the form of temporary
versus permanent downsizing or voluntary versus compulsory downsizing during at least one of the quarters of 2020. We also
considered reductions in the working week and salary reductions to be compulsory reductions in labor costs. Some firms may have
engaged in more than one way of reducing labor costs in the same quarter, including both voluntary and compulsory redundancies.
In such cases, in the regression analysis, we considered the worst form of reducing labor costs that a firm used. As mentioned in
Section 3.1, after identifying the Form 8-K filings containing some or all of the keywords for labor cost reductions, we encode the
following indicator variables to capture the occurrence of different types of downsizing:
• Labor cost reductions (any of the following ways of reducing labor costs),
• Temporary (temporary downsizing measures, such as furloughing),
• Permanent (permanent downsizing measures, such as early retirement),
• Voluntary (voluntary leave),
• Compulsory (compulsory leave),
• Shortened (shortened workweek program),
• Reduced salary (a salary cut for all staff), and
• Reduced salary board (a salary cut for the board only).21
With the help of excerpts from seven Form 8-Ks, Table B.1 in Appendix B illustrates how the encoding of the variables was
operationalized. We then aggregate these indicator variables at the annual level for calendar year 2020.22
As stated in Section 2, the retail, leisure, hospitality, and travel industries suffered the most from the pandemic; they were
unable to operate during part of 2020 because of lockdowns and travel restrictions (Chetty et al., 2022; Bartik et al., 2020). Firms
in these industries typically did not have the option to avoid reductions in labor costs. Nevertheless, the CEOs of such firms may
still have had a choice among the different ways of reducing labor costs. Notably, employees should prefer temporary downsizing
over permanent downsizing. They should also prefer voluntary redundancies over compulsory redundancies. Having data on the
different ways of reducing labor costs enables us to assess the severity of the labor cost reductions. Finally, they should prefer
salary reductions for board directors only to salary reductions for the entire staff. In line with our main hypothesis, we expect that
conservative CEOs will opt for more severe forms of labor cost reductions. To test whether the CEOs’ political ideology affected
18 Note that we do not have any cases of firms repurchasing shares while their dividend is below expectations.
19 The reader should note that in earlier versions of the paper, we also considered the reaction ‘‘employee pain and shareholder joy’’ in our analysis. This
reaction consisted of reducing labor costs while exceeding dividend expectations. However, given the relatively small number of observations for this reaction,
we decided to merge it with ‘‘employee pain.’’ This is different from the reaction ‘‘no employee pain and shareholder joy’’, which we still consider.
20 Put differently, for each firm, we focus on the worst 2020 dividend reaction to the pandemic.
21 The encoding of the various labor cost reduction variables was conducted by two members of the data-encoding team. Any discrepancies in the encoding
between the two members were then resolved by the third member of the team.
22 Earlier versions of the paper also included the variable labor cost reduction date for each quarter, which indicates the earliest date when reductions in labor
costs occurred in 2020. This date is compared with the date when the DPS changed. For only a few firms, the dividend (labor cost reduction) decision preceded
the labor cost reduction (dividend) decision. Hence, in general, there was no timing difference between the two decisions.
6
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
the severity of the reductions for those firms engaging in such reductions, we rerun Eq. (1), this time distinguishing between firms
engaging in permanent labor cost reductions and those engaging in temporary labor cost reductions.
For both versions of Eq. (1), our key variable of interest is CEO political ideology. CEO political ideology is measured in two
different ways. First, we use CEO conservatism. Considering the lifetime political donations of the CEO up to and including calendar
year 2020, but ignoring the donations made in the calendar year before U.S. presidential elections,23 this alternative measure is
based on the total amount that the CEO donated to the Republican Party divided by the sum of the total amounts donated to
both the Republican Party and the Democratic Party. Second, we use a set of four indicator variables measuring the political
leanings of CEOs: Conservatives, CEOs who donated to the Republican Party only; Liberals, CEOs who donated to the Democratic
Party only; Nonpartisans, CEOs who donated to both the Democratic and Republican parties; and Zerodonations, CEOs who did not
make any donations. It should be noted that there are limitations to these indicator variables. First, we end up with relatively few
conservative CEOs (64). Second, a CEO who donated $100,000 to the GOP and $500 to the Democrats would be categorized as
nonpartisan, similar to another CEO who donated $100,000 to the Democrats and $500 to the Republicans. Given these limitations,
the descriptive tables that follow use the index rather than the four CEO categories to identify subsamples based on CEO political
ideology. Nevertheless, in the regression analysis, we use both and find qualitatively similar results.
While not a key variable of interest, we nevertheless consider whether a firm reported a loss in any of the four quarters of 2020
as a key determinant of its type of reaction to the pandemic. Hence, we include 𝐿𝑜𝑠𝑠, an indicator variable set to one if earnings per
share (EPS) in at least one of the four quarters of 2020 are negative, and zero otherwise, in our regressions. The use of this indicator
variable is motivated by DeAngelo et al. (1992), who found that a loss – independent of its size – was a necessary condition for
firms on the New York Stock Exchange (NYSE) to reduce their dividend during 1980–1985. The majority of firms on the NYSE with
a loss during that period reduced their dividend, compared with only 1% of firms without a loss. Interestingly, their study also
implies that the percentage of firms that did not cut their dividend in the wake of a loss is just below a majority. Hence, while a loss
is a necessary condition to cut the dividend, it is not a sufficient condition. We also interact 𝐶 𝐸 𝑂 𝑝𝑜𝑙𝑖𝑡𝑖𝑐 𝑎𝑙 𝑖𝑑 𝑒𝑜𝑙𝑜𝑔 𝑦 with the 𝐿𝑜𝑠𝑠
indicator variable. This interaction enables us to identify possible differences in behavior based on CEO political ideology between
loss-making firms and the remaining ones.
3.2.2. Control variables
Our first two control variables account for firm behavior during ‘‘normal times’’. These variables are the standard deviation
of the percentage change in the number of employees (𝐸 𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑆 𝑡. 𝐷𝑒𝑣. ) and the standard deviation of the dividend surprise
(𝐷𝑖𝑣. 𝑆 𝑢𝑟𝑝𝑟𝑖𝑠𝑒 𝑆 𝑡. 𝐷𝑒𝑣. ) for the period of 2015–2019. To ensure representative and meaningful values for the standard deviations,
we only computed the standard deviation in question if there were at least three observations for the percentage change in the
number of employees or the dividend surprise for the period of 2015–2019. As sufficient data on the number of employees (the
dividend surprise) were available for only 398 (437) firms, we set the standard deviations to zero for firms with insufficient data.
Next, we include measures that have been shown to explain changes in dividends. In line with DeAngelo et al. (1992), we include
the 𝐿𝑜𝑠𝑠 indicator variable as mentioned above. As per Lintner’s (1956) dividend model, we control for 𝐸 𝑃 𝑆 and 𝛥 𝐸 𝑃 𝑆, which
is EPS for calendar year 2020, and the difference between EPS for calendar year 2020 and EPS for calendar year 2019. Although
the following variable has not yet been used to explain changes in dividends, given the way we measure dividend changes (i.e., by
comparing the actual DPS in 2020 with the expected DPS), we also include 𝐸 𝑃 𝑆 𝑠𝑢𝑟𝑝𝑟𝑖𝑠𝑒 (i.e., the difference between the actual 𝐸 𝑃 𝑆
for the fiscal year 2020 and the expected 𝐸 𝑃 𝑆 for the same year). We also consider stock repurchase behavior during the calendar
years 𝑡, where 𝑡 = 2019, 2020. Hence, we include the three following variables: 𝑅𝑒𝑝𝑢𝑟𝑐 ℎ𝑎𝑠𝑒𝑠𝑡 is an indicator variable set to one if
there were stock repurchases in calendar year 𝑡 and zero otherwise; 𝑆 ℎ𝑎𝑟𝑒𝑠 𝑟𝑒𝑝𝑢𝑟𝑐 ℎ𝑎𝑠𝑒𝑑𝑡 is the ratio of total shares repurchased in
year 𝑡 to common shares outstanding in year 𝑡 − 1; and 𝑉 𝑎𝑙𝑢𝑒 𝑠ℎ𝑎𝑟𝑒𝑠 𝑟𝑒𝑝𝑢𝑟𝑐 ℎ𝑎𝑠𝑒𝑑𝑡 is the product of total shares repurchased in year
𝑡 and the average price per repurchased share for the same year.
Our next two control variables are measures that may affect the likelihood of labor cost reductions. We include the two key
variables that are included in the classic labor demand equation (see Nickell, 1984): wage costs and labor productivity. These two
variables adjust for labor rigidity; we expect labor productivity and wage costs to be greater in firms with high levels of firm-
specific human capital (see Nickell, 1984 for further details). 𝑊 𝑎𝑔 𝑒 𝑐 𝑜𝑠𝑡𝑠2019 is the natural logarithm of wages, and 𝑃 𝑟𝑜𝑑 𝑢𝑐 𝑡𝑖𝑣𝑖𝑡𝑦2019
is measured by the natural logarithm of sales, both relating to the fiscal year 2019. One limitation of the Compustat database is the
lack of data on labor costs, which are only available for 12% of the firm-year observations in our sample. To deal with this lack of
data, we complement the firm-year observations available in Compustat with data obtained from the Annual Survey of Manufactures
(ASM) conducted by the U.S. Census Bureau. Following Donangelo et al. (2019), we estimate the total labor costs to the employer,
using data at the two-digit NAICS level, as the logarithm of the industry average for the sum of salaries and wages plus additional
costs (for more details, see Appendix D). Hence, we assume that firms with missing labor costs have the same labor costs as the
average firm in their industry.24
We include 𝐼 𝑛𝑠𝑡𝑖𝑡𝑢𝑡𝑖𝑜𝑛𝑎𝑙 𝑜𝑤𝑛𝑒𝑟𝑠ℎ𝑖𝑝 𝑟𝑎𝑡𝑖𝑜2019 as the ratio of institutional ownership to the total number of shares outstanding,
based on the calendar year end for 2019. Furthermore, the indicator variable 𝐶 𝐴𝑅𝐸 𝑆 is set to one if the firm applied for assistance
23 Donations made during the year prior to a presidential election tend to be driven by opportunism rather than ideology. For details, see Bayat and Goergen
(2022).
24 Our results to do not change qualitatively when excluding this variable from the regressions.
7
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
under the CARES Act during the calendar year 2020 and zero otherwise. In turn, 𝑅𝑒𝑑 𝑠𝑡𝑎𝑡𝑒 is an indicator variable that is set to one
if the firm’s headquarters are located in a state where a majority of voters voted for the Republicans in the 2019 elections.25
Also considered are several variables measuring CEO characteristics, including 𝐶 𝐸 𝑂 𝑎𝑔 𝑒, and a number of indicator variables
set to one if the CEO is male (𝐶 𝐸 𝑂 𝑔 𝑒𝑛𝑑 𝑒𝑟); the CEO also acts as the chair of the board of directors (𝐶 𝐸 𝑂 𝑑 𝑢𝑎𝑙𝑖𝑡𝑦); and the CEO
has an MBA (𝑀 𝐵 𝐴), a PhD (𝑃 ℎ𝐷), a degree from an Ivy League university (𝐼 𝑣𝑦 𝐿𝑒𝑎𝑔 𝑢𝑒), or a professional accounting qualification
(𝑃 𝑟𝑜𝑓 𝑒𝑠𝑠𝑖𝑜𝑛𝑎𝑙 𝑞 𝑢𝑎𝑙𝑖𝑓 𝑖𝑐 𝑎𝑡𝑖𝑜𝑛). We also control for CEO share ownership.26
Finally, we also utilize a set of control variables related to firm characteristics — namely, 𝐹 𝑖𝑟𝑚 𝑠𝑖𝑧𝑒, measured by the natural
logarithm of total assets; 𝐹 𝑖𝑟𝑚 𝑎𝑔 𝑒, measured by the natural logarithm of the number of years that the firm has been included in
CRSP up to and including the year 2019; 𝑇 𝑜𝑏𝑖𝑛′ 𝑠 𝑄, defined as the ratio of total assets plus the product of common shares outstanding
at the end of fiscal year 2019 close price minus the book value of common equity over total assets; 𝐶 𝑎𝑠ℎ ℎ𝑜𝑙𝑑 𝑖𝑛𝑔 𝑠, measured as
the ratio of cash and short-term investments over total assets; the 𝐸 𝑆 𝐺 𝑠𝑐 𝑜𝑟𝑒 from ASSET4/Refinitiv ESG, which ranges from 0
to 100; 𝐶 𝑟𝑒𝑑 𝑖𝑡 𝑟𝑎𝑡𝑖𝑛𝑔, an indicator variable set to one for firms with a credit rating of investment grade (i.e., a rating of BBB- or
better) and zero otherwise; 𝑀 𝑎𝑟𝑘𝑒𝑡-𝑡𝑜-𝑏𝑜𝑜𝑘, measured as the ratio of total market value of equity over the book value of equity;
𝑅 & 𝐷, measured as research and development expenses; 𝐹 𝑜𝑟𝑒𝑖𝑔 𝑛 𝑠𝑎𝑙𝑒𝑠, which is equal to the ratio of foreign sales over foreign and
domestic sales; 𝐶 𝑎𝑝𝑖𝑡𝑎𝑙 𝑒𝑥𝑝𝑒𝑛𝑑 𝑖𝑡𝑢𝑟𝑒𝑠, defined as the ratio of capital expenditures over total assets; and 𝐿𝑒𝑣𝑒𝑟𝑎𝑔 𝑒, defined as the book
value of long-term debt over the book value of total assets. Detailed variable definitions can be found in Appendix D.
4. Empirical analysis
4.1. Descriptive statistics
Table 1 reports summary statistics for the variables used in this study. Table 2 documents the correlation coefficients between
these variables. Both tables are based on observations for the calendar year 2020 unless otherwise specified.
Table 1 suggests that for 43.1% of the firms, at least one of the actual quarterly dividends per share in calendar year 2020 is
below the expected quarterly dividend. The percentage of firms reducing their labor costs is lower, at 28.2%. In addition, most
labor cost reductions tend to be compulsory downsizing (not tabulated). Further, 8.3% of firms also reduced the salaries of the
entire workforce, and 13.7% of the firms reduced the remuneration of the board of directors (not tabulated). Just under 14% of
firms have a CEO classed as conservative, whereas 6.8% of firms have a liberal CEO. Nonpartisans – that is, CEOs who donated to
both the Democratic and Republican parties – make up 24.3% of the observations. At first sight, the low percentage of conservative
CEOs might appear surprising. However, the reader should remember that the definition of a conservative CEO, which underlies
the 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠 indicator variable, is extremely stringent. Indeed, we only consider a CEO to be conservative if the CEO has only
ever donated to the Republican Party. This would exclude a CEO whose donations were mostly to the Republican Party, with the
odd smaller donation to the Democrats. Nevertheless, in line with expectations, the average for 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 is above 0.5 at
0.563, suggesting that the average CEO is conservative. At the same time, the median is exactly 0.5. Importantly, while we use two
different sets of measures for CEO political ideology, resulting in very different distributions of conservative CEOs and other CEOs,
our regression results from using the two alternative sets of measures are not qualitatively different. This confirms the robustness
of our key results. Furthermore, 34.4% of the firms reported a loss; on average, however, actual earnings per share were greater
than expected (see 𝐸 𝑃 𝑆 𝑠𝑢𝑟𝑝𝑟𝑖𝑠𝑒). Moreover, there was a drop in the ratio of shares repurchased in 2020 compared to 2019. The
same pattern applies to the value of shares repurchased. About 10% of the firms successfully applied for support under the CARES
Act. Finally, average CEO age is approximately 59 years, 37% of CEOs also act as chair of the board of directors, and slightly above
5% of CEOs are female. These descriptive statistics for the CEO characteristics are in line with those reported in extant studies (see
e.g. De Angelis and Grinstein, 2020). Note that we confirm the continuation of the downward trend in the percentage of firms with
CEO duality, as highlighted by Graham et al. (2020).
Table 2 reports the correlation matrix. The table documents a significant and positive association between 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚
and the actual dividend being above expectations. The same pattern is observed for the indicator variable 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠. In contrast,
the correlation between the dividend exceeding expectations on the one side and the indicator variables 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠, 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠, and
𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠 on the other is either negative or insignificant. Interestingly, the indicator variable 𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠 also suggests that
CEOs without political donations are more likely to pay a dividend below the expected one. Nevertheless, no association between
the measures of political ideology and labor cost reductions is found. Finally, there is no significant correlation between the 𝐶 𝐴𝑅𝐸 𝑆
indicator variable on the one hand and 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 or any of the four indicator variables of CEO political ideology on the
other. Hence, there is no evidence that CEO political ideology affected the likelihood of a firm applying for assistance under the
CARES Act.
25 It would not make sense to base this indicator variable on the 2020 presidential elections, as the way voters voted may have been a reflection of their
satisfaction with how the Trump administration had managed the pandemic.
26 When we interact CEO share ownership with the loss indicator variable (not tabulated), we do not find that the coefficient on Loss * CEO share ownership
is significant in any of these regressions.
8
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 1
Summary statistics.
Statistic Mean Median St. Dev. No. Obs.
Actual Div. Above Expectation 0.312 0 0.464 462
Actual Div. Below Expectation 0.431 0 0.496 462
Actual Div. Equals Expectation 0.257 0 0.437 462
Labor Cost Reductions 0.282 0 0.450 481
Conservatives 0.139 0 0.346 481
Liberals 0.068 0 0.253 481
Nonpartisans 0.243 0 0.429 481
Zerodonations 0.548 1 0.498 481
CEO Conservatism 0.563 0.500 0.273 481
Employee St. Dev. 8.087 4.447 12.106 481
Div. Surprise St. Dev. 0.126 0.024 0.253 481
EPS 1.386 0.857 11.50 479
𝛥EPS −0.222 0.010 2.592 481
Loss 0.344 0 0.475 479
EPS Surprise 0.698 0.396 1.673 481
Repurchases2019 0.844 1 0.363 481
Repurchases2020 0.804 1 0.396 481
Shares Repurchased2019 0.027 0.016 0.033 481
Shares Repurchased2020 0.014 0.007 0.023 479
Value Shares Repurchased2019 1933.634 365.461 5972.825 481
Value Shares Repurchased2020 1383.461 150.032 10,550.810 479
CARES 0.099 0 0.300 481
Institutional Ownership Ratio2019 2.350 2.430 0.619 479
Productivity2019 8.070 7.922 1.263 481
Industry Labor Expense2019 5.079 0 8.516 481
Red State 0.419 0 0.494 481
CEO Age 58.690 59 6.620 481
CEO Gender 0.946 1 0.226 481
CEO Duality 0.370 0 0.483 481
MBA 0.401 0 0.491 481
PhD 0.033 0 0.180 481
Ivy League 0.104 0 0.306 481
Professional Qualification 0.060 0 0.238 481
CEO Share Ownership 0.251 0.101 0.424 481
Tobin’s Q 2.440 1.797 1.945 481
Tobin’s Q2020 2.601 1.769 2.340 461
Tobin’s Q2021 2.717 1.900 2.563 481
Tobin’s Q2022 2.449 1.791 1.939 443
ROA2020 0.033 0.029 0.087 442
ROA2021 0.063 0.050 0.072 436
ROA2022 0.068 0.043 0.078 388
ROE2020 0.452 0.112 6.213 461
ROE2021 −0.422 0.159 9.627 455
ROE2022 0.096 0.143 4.752 405
Cash Holdings 0.106 0.059 0.127 481
ESG Score 44.544 57.150 31.908 481
Credit Rating 0.097 0 0.297 481
Firm Size2019 9.925 9.915 1.897 481
Firm Age2019 3.386 3.508 0.892 481
Market to Book 1.129 2.516 73.187 481
R&D2019 684.941 0 2678.871 481
Foreign Sales2019 0.273 0.241 0.267 481
Capital Expenditures2019 0.033 0.0232 0.034 481
Leverage2019 0.312 0.309 0.201 481
China 0.291 0 0.455 481
This table reports the mean, median, and standard deviation for the S&P 500 firms headquartered in the U.S. for the COVID-19 period, i.e., the calendar year
2020. The rightmost column reports the number of observations for each of the variables in our sample. We also report the repurchases indicator, the shares
repurchased, the value of shares repurchased, productivity, and industry labor expense for 2019.
4.2. Univariate analysis
Table 3 reports the types of reaction of the S&P 500 firms to the pandemic. The level of the observation is the firm. The
table distinguishes between firms with conservative CEOs (i.e, firms for which 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 exceeds 0.5) and all other firms
(i.e., firms for which 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 is equal to or below 0.5). The main numbers in the table are the percentages of firms
for a given level of 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 that engage in a specific combination of labor cost reduction and dividend reaction to the
pandemic, whereas the numbers in parentheses are the numbers of firms adopting a specific combination of labor cost reductions
and dividend decision.
9
A. Bayat et al.
Table 2
Correlation matrix.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
1. Actual Div. Above Expectation 1
2. Actual Div. Below Expectation −0.59 1
3. Actual Div. Equals Expectation −0.40 −0.51 1
4. Labor Cost Reductions −0.11 0.10 −0.01 1
5. Conservatives 0.14 −0.12 −0.01 −0.04 1
6. Liberals 0.08 −0.08 0.01 0.01 −0.11 1
7. Nonpartisans −0.04 −0.01 0.04 0.04 −0.23 −0.15 1
8. Zerodonations −0.10 0.12 −0.03 −0.01 −0.44 −0.30 −0.63 1
9. CEO Conservatism 0.11 −0.05 −0.06 −0.03 0.61 −0.54 0.13 −0.26 1
10. EPS −0.02 −0.04 0.06 −0.08 −0.03 −0.02 −0.03 0.05 −0.02 1
11. 𝛥EPS 0.09 0.01 −0.10 −0.10 −0.07 −0.01 −0.03 0.07 0.01 0.19 1
12. Loss −0.12 0.06 0.06 0.29 0.06 −0.01 0.05 −0.08 0.07 −0.15 −0.13 1
13. EPS Surprise −0.05 −0.05 0.11 −0.06 0.01 −0.08 −0.02 0.05 0.01 0.16 0.13 −0.16 1
14. Repurchases2020 −0.01 −0.04 0.05 0.03 −0.01 −0.09 −0.01 0.07 0.01 0.04 −0.02 −0.10 0.06 1
15. Repurchases2019 −0.02 −0.03 0.06 0.04 0.04 −0.09 −0.06 0.07 0.01 0.03 −0.02 −0.01 0.03 0.70 1
16. Share Repurchased2020 −0.02 −0.14 0.17 −0.09 −0.04 −0.02 −0.01 0.04 −0.07 0.07 −0.01 −0.14 0.15 0.30 0.18 1
17. Share Repurchased2019 −0.05 −0.02 0.08 0.06 0.03 −0.05 −0.02 0.02 −0.01 0.06 −0.04 −0.04 0.06 0.30 0.31 0.47 1
18. Value Share Repurchased2020 −0.04 −0.04 0.09 −0.06 −0.02 −0.01 0.02 −0.01 −0.04 0.02 −0.02 −0.06 0.06 0.06 −0.07 0.45 0.03 1
19. Value Share Repurchased2019 −0.05 0.08 −0.04 −0.06 0.04 −0.04 0.16 −0.15 −0.01 0.01 −0.04 −0.09 0.02 0.13 0.14 0.12 0.33 0.20 1
20. CARES 0.02 −0.04 0.03 0.19 0.05 −0.01 0.01 −0.03 0.01 −0.04 −0.10 0.15 0.06 0.04 0.05 0.01 0.04 −0.03 −0.06 1
21. Institutional Ownership Ratio2019 −0.01 0.02 −0.01 0.06 −0.01 0.07 −0.07 0.03 −0.04 0.01 0.03 −0.04 −0.07 0.04 −0.04 −0.08 −0.01 −0.21 −0.13 −0.02 1
22. Productivity2019 −0.01 0.01 −0.01 0.04 0.02 −0.08 0.10 −0.06 0.03 0.03 −0.14 0.13 0.14 0.12 0.15 0.09 0.13 0.19 0.33 0.13 −0.33 1
23. Wage Cost2019 0.01 −0.01 0.01 −0.09 −0.05 −0.08 0.07 0.01 0.04 0.00 −0.02 −0.11 0.05 0.06 0.07 0.00 0.03 −0.02 0.12 −0.04 −0.02 0.01 1
24. Red State 0.01 0.02 −0.02 0.00 0.16 −0.10 −0.04 −0.02 0.20 −0.06 −0.02 0.10 −0.05 −0.04 −0.03 −0.07 0.01 −0.06 −0.03 0.03 0.02 0.04 −0.10 1
25. CEO Age 0.01 0.05 −0.07 0.09 −0.03 0.03 0.08 −0.06 0.01 0.03 −0.03 0.12 −0.03 0.01 −0.02 −0.03 −0.06 −0.04 0.01 0.07 −0.02 0.06 0.01 0.08
26. CEO Gender 0.02 −0.09 0.08 −0.03 0.04 −0.04 −0.06 0.04 0.05 0.03 −0.03 −0.02 0.04 −0.05 −0.05 0.01 −0.02 0.01 0.01 −0.07 0.03 −0.05 0.07 −0.02
27. CEO Duality −0.02 0.05 −0.04 0.07 −0.04 −0.04 0.22 −0.14 0.11 −0.04 0.01 0.01 0.01 0.05 0.04 0.04 −0.01 −0.03 −0.04 0.05 −0.03 0.00 −0.01 0.01
28. MBA 0.02 0.01 −0.03 0.02 0.11 −0.04 0.03 −0.08 0.09 0.09 0.02 −0.08 0.03 0.06 0.01 0.04 0.01 0.01 −0.01 −0.04 0.04 −0.01 0.09 0.02
29. PhD −0.05 0.05 −0.01 −0.12 −0.07 0.04 0.03 0.01 −0.05 0.02 −0.01 −0.03 0.06 0.01 0.05 0.01 −0.06 −0.01 −0.01 −0.06 0.03 −0.02 0.06 −0.11
30. Ivy League −0.09 0.03 0.06 −0.01 −0.05 0.01 0.01 0.03 −0.03 0.01 0.03 −0.01 0.12 0.06 0.03 0.06 −0.04 0.00 −0.03 0.01 −0.03 −0.03 0.05 −0.07
31. Professional Qualification 0.02 −0.01 −0.02 0.01 0.03 −0.03 0.03 −0.03 0.02 −0.01 −0.02 0.07 −0.01 −0.04 −0.04 −0.04 −0.08 −0.02 −0.06 0.10 0.07 0.01 −0.07 −0.01
32. CEO Share Ownership −0.07 −0.06 0.14 0.04 0.01 −0.01 0.01 −0.01 0.00 0.12 0.04 0.03 0.14 0.00 −0.02 0.15 −0.05 −0.01 −0.08 0.04 −0.05 −0.08 −0.06 −0.01
33. Tobin’s Q −0.01 −0.07 0.08 −0.09 −0.04 0.04 −0.04 0.04 −0.04 0.08 0.03 −0.25 0.04 0.04 0.02 −0.03 −0.10 −0.01 −0.01 −0.12 −0.03 −0.25 0.16 −0.19
34. Cash Holdings −0.05 −0.09 0.15 −0.11 −0.11 0.17 −0.07 0.05 −0.17 0.09 0.01 −0.05 0.05 0.04 0.07 0.11 0.10 0.03 0.16 −0.07 −0.02 −0.10 0.08 −0.29
10 35. ESG Score 0.10 0.05 −0.18 −0.04 −0.02 0.02 −0.05 0.04 −0.01 −0.01 0.12 −0.08 0.06 −0.06 −0.03 −0.11 −0.02 −0.05 0.08 −0.01 0.03 0.07 0.17 −0.04
36. Credit Rating −0.04 0.02 0.01 −0.03 0.02 −0.03 −0.01 0.01 0.03 −0.01 −0.01 0.01 −0.01 0.02 0.06 −0.04 −0.01 −0.01 0.02 0.03 −0.05 0.18 0.01 0.03
37. Firm Size2019 0.03 0.06 −0.10 −0.10 −0.02 −0.07 0.10 −0.03 0.03 −0.04 −0.02 0.06 0.07 0.02 −0.02 0.03 0.06 0.09 0.25 −0.02 −0.16 0.42 0.08 −0.01
38. Firm Age2019 0.13 0.03 −0.17 0.04 0.04 −0.03 0.01 −0.02 0.07 0.01 0.01 −0.06 −0.01 −0.02 0.01 −0.05 0.01 −0.05 0.07 0.05 0.07 0.11 0.08 0.05
39. Market to Book −0.05 0.04 0.01 0.02 0.01 −0.02 0.01 0.01 0.01 0.01 −0.01 0.01 0.01 −0.01 −0.02 −0.01 0.01 0.01 0.01 −0.01 0.01 −0.04 −0.05 0.01
40. R&D2019 −0.03 −0.02 0.07 −0.08 0.01 0.03 0.07 −0.08 −0.04 0.04 0.03 −0.04 0.46 −0.01 −0.01 0.03 0.01 0.13 0.25 −0.06 −0.17 0.28 0.04 −0.13
41. Foreign Sales2019 −0.01 −0.04 0.07 0.01 −0.01 0.12 −0.07 0.01 −0.08 −0.05 −0.02 −0.03 −0.01 0.01 0.06 −0.02 0.01 −0.02 0.06 −0.18 −0.01 0.01 0.08 −0.15
42. Capital Expenditures2019 0.01 −0.02 0.01 0.10 0.14 0.01 −0.04 −0.06 0.07 −0.09 −0.07 0.16 −0.01 −0.11 −0.09 −0.10 −0.10 0.04 −0.08 0.07 −0.10 0.05 −0.13 0.21
43. Leverage2019 −0.06 0.16 −0.10 0.15 0.03 −0.03 0.02 −0.02 0.06 −0.03 −0.07 0.10 −0.09 −0.06 −0.07 −0.05 −0.03 0.02 −0.03 0.05 0.04 −0.02 −0.09 0.08
44. Employee St. Dev. 0.02 −0.06 0.04 0.04 0.01 0.01 −0.01 0.01 −0.01 −0.02 −0.01 0.03 0.01 −0.05 −0.09 0.07 0.01 0.01 −0.02 0.06 0.04 −0.02 −0.03 0.06
45. Div. Surprise St. Dev. 0.09 0.02 −0.13 0.03 0.05 −0.01 0.01 −0.03 0.13 0.57 0.82 0.95 0.26 0.75 0.01 −0.04 −0.03 −0.02 −0.03 0.06 −0.05 0.03 −0.02 0.02
46. Tobin’s Q2020 −0.03 −0.11 0.17 −0.09 −0.03 0.04 −0.03 0.03 −0.04 0.05 0.04 −0.28 0.06 0.03 0.01 −0.03 −0.11 0.00 −0.00 −0.12 −0.04 −0.25 0.14 −0.19
47. Tobin’s Q2021 −0.02 −0.09 0.12 −0.06 −0.01 0.02 −0.02 0.02 −0.01 0.07 0.03 −0.27 0.07 0.04 0.00 −0.03 −0.09 0.00 0.01 −0.11 −0.02 −0.23 0.13 −0.16
48. Tobin’s Q2022 −0.00 −0.07 0.09 −0.07 −0.01 0.02 −0.03 0.02 0.00 0.06 0.04 −0.26 0.05 0.09 0.06 −0.02 −0.06 0.01 0.03 −0.11 −0.04 −0.19 0.14 −0.14
49. ROA2020 0.04 −0.07 0.04 −0.20 −0.07 0.04 −0.01 0.04 −0.07 0.20 0.22 −0.53 0.18 0.07 0.06 0.16 0.06 0.03 0.05 −0.15 0.01 −0.12 0.18 −0.11
50. ROA2021 0.01 −0.03 0.02 −0.05 0.04 −0.01 −0.07 0.04 0.02 0.17 0.18 −0.28 0.19 0.16 0.13 0.12 0.14 0.01 0.04 −0.07 −0.04 −0.02 0.09 −0.08
51. ROA2022 0.02 −0.01 −0.01 −0.04 0.07 0.01 −0.06 −0.00 0.03 0.19 0.12 −0.12 0.09 0.16 0.12 0.01 0.09 0.00 0.06 −0.03 −0.06 0.01 0.03 −0.01
52. ROE2020 0.06 0.01 −0.07 −0.10 −0.05 0.04 0.02 −0.01 −0.03 0.14 0.04 −0.39 0.12 0.08 0.02 0.16 0.04 0.03 0.04 −0.10 0.04 −0.08 0.10 −0.10
Journal of Corporate Finance 90 (2025) 102692
53. ROE2021 0.05 0.03 −0.08 −0.11 0.07 −0.07 −0.04 0.02 0.07 0.09 0.25 −0.14 0.11 0.10 0.06 0.17 0.23 0.01 0.09 −0.12 −0.01 0.10 0.08 0.05
54. ROE2022 −0.00 0.08 −0.08 0.01 0.05 −0.08 −0.05 0.05 0.07 0.06 0.13 −0.05 0.09 0.13 0.10 0.05 0.21 −0.02 0.06 0.06 0.00 0.13 −0.00 0.03
55. China −0.05 −0.01 0.07 0.04 −0.05 0.05 −0.01 0.02 −0.08 −0.01 −0.01 −0.06 0.01 −0.01 0.04 −0.01 0.02 −0.01 0.07 −0.07 0.02 −0.02 0.07 −0.15
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
25. CEO Age 1
26. CEO Gender 0.07 1
27. CEO Duality 0.24 0.06 1
28. MBA −0.02 −0.04 −0.01 1
29. PhD −0.01 0.04 −0.06 −0.10 1
30. Ivy League 0.00 0.06 0.03 0.03 0.05 1
31. Professional Qualification 0.05 0.06 0.01 −0.01 0.01 −0.06 1
32. CEO Share Ownership 0.24 0.01 0.15 0.02 0.06 0.17 −0.04 1
33. Tobin’s Q −0.01 −0.01 −0.06 0.07 0.05 0.01 −0.06 0.08 1
34. Cash Holdings −0.01 −0.04 −0.01 −0.02 0.05 0.01 −0.07 0.14 0.45 1
35. ESG Score −0.02 −0.03 −0.05 0.04 0.03 −0.01 −0.07 −0.08 0.06 0.01 1
36. Credit Rating 0.02 −0.04 0.03 −0.08 −0.02 −0.02 −0.01 −0.03 0.03 0.04 0.04 1
37. Firm Size2019 0.17 0.03 0.01 −0.02 −0.09 0.01 0.02 −0.14 −0.18 −0.04 0.14 0.12 1
38. Firm Age2019 0.13 0.03 0.03 −0.01 −0.02 −0.08 0.06 −0.10 −0.05 −0.09 0.31 0.09 0.09 1
(continued on next page)
A. Bayat et al.
Table 2 (continued).
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55
39. Market to Book −0.02 −0.01 0.01 −0.08 0.01 0.01 −0.01 −0.01 0.01 −0.06 0.07 0.01 0.01 −0.02 1
40. R&D2019 −0.04 0.01 −0.04 0.04 0.04 0.08 −0.04 0.07 0.10 0.24 0.11 0.10 0.19 0.01 0.02 1
41. Foreign Sales2019 −0.01 0.07 −0.04 0.06 0.12 −0.05 −0.12 −0.03 0.24 0.33 0.14 0.09 0.01 −0.03 −0.05 0.20 1
42. Capital Expenditures2019 0.02 −0.03 −0.09 0.01 −0.02 −0.03 −0.01 0.01 0.03 −0.12 0.01 0.02 0.03 0.05 −0.05 0.06 −0.02 1
43. Leverage2019 −0.02 −0.02 −0.08 0.02 0.01 0.08 0.05 −0.04 0.12 −0.13 0.03 0.04 −0.09 0.02 −0.04 −0.04 −0.01 0.18 1
44. Employee St. Dev. 0.02 0.07 −0.01 0.06 −0.06 0.08 0.01 −0.01 −0.05 −0.08 −0.13 −0.02 0.08 −0.05 0.02 0.01 0.08 0.01 0.08 1
45. Div. Surprise St. Dev. 0.02 −0.03 −0.04 −0.04 0.04 −0.01 −0.02 −0.04 −0.03 −0.02 −0.04 −0.06 0.15 0.02 0.03 −0.0 −0.01 0.04 0.09 0.07 1
46. Tobin’s Q2020 −0.05 −0.01 −0.04 0.07 0.11 −0.01 −0.07 0.11 0.96 0.42 0.03 0.02 −0.44 −0.11 0.02 0.12 0.23 0.02 0.08 −0.06 −0.03 1
11 47. Tobin’s Q2021 0.01 −0.03 −0.05 0.10 0.04 −0.02 −0.06 0.13 0.90 0.40 0.05 0.04 −0.18 −0.05 0.03 0.12 0.26 0.04 0.12 −0.04 −0.01 0.95 1
48. Tobin’s Q2022 −0.03 0.01 −0.04 0.08 0.10 −0.03 −0.07 0.12 0.83 0.40 0.09 0.11 −0.39 −0.09 −0.01 0.10 0.23 0.02 0.18 −0.09 0.01 0.85 0.89 1
49. ROA2020 −0.10 −0.03 −0.03 0.07 0.08 0.03 −0.09 0.01 0.52 0.28 0.13 0.01 −0.25 0.05 −0.01 0.15 0.16 −0.21 −0.08 −0.12 −0.02 0.53 0.53 0.49 1
50. ROA2021 −0.13 −0.07 −0.05 0.08 0.04 −0.01 −0.13 0.01 0.48 0.27 0.12 0.06 −0.31 0.03 0.02 0.20 0.26 0.12 0.05 −0.06 0.05 0.49 0.55 0.53 0.67 1
51. ROA2022 −0.04 −0.02 −0.04 0.06 0.02 −0.01 −0.11 0.01 0.33 0.21 0.06 0.04 −0.28 0.03 −0.01 0.13 0.21 0.26 0.11 −0.04 0.06 0.35 0.45 0.54 0.39 0.69 1
52. ROE2020 −0.04 −0.02 0.05 0.00 0.03 0.04 −0.01 0.03 0.10 0.06 0.07 −0.05 0.01 −0.00 0.02 0.04 −0.10 −0.27 −0.16 −0.08 −0.09 0.09 0.09 0.07 0.55 0.19 −0.03 1
53. ROE2021 −0.02 −0.02 0.03 0.01 −0.00 0.01 −0.02 0.00 −0.11 −0.03 0.01 0.00 0.23 −0.01 0.01 0.02 −0.10 −0.13 −0.17 0.05 −0.00 −0.11 −0.13 −0.11 0.24 0.26 0.12 0.24 1
54. ROE2022 −0.07 −0.07 0.00 −0.04 −0.07 −0.01 −0.02 −0.06 −0.12 −0.13 0.03 −0.03 0.12 0.03 −0.01 −0.02 −0.11 0.05 −0.07 0.01 0.03 −0.11 −0.10 −0.08 0.01 0.21 0.44 0.08 0.41 1
55. China 0.01 0.01 −0.01 0.06 0.01 −0.09 −0.05 −0.04 0.21 0.24 0.07 0.06 −0.01 0.01 0.06 0.06 0.55 −0.05 −0.05 0.06 0.02 0.20 0.21 0.17 0.17 0.16 0.06 0.06 −0.04 −0.04 1
The table reports the Pearson correlation coefficients for the COVID-19 period – that is, the calendar year 2020 – unless otherwise specified. Numbers in bold indicate correlation coefficients that are statistically significant at the 10% level. Appendix D provides the
definitions of all variables.
Journal of Corporate Finance 90 (2025) 102692
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 3
Reaction to pandemic by CEO conservatism.
Firms with CEO conservatism > 0.5 Firms with CEO conservatism ≤ 0.5
Labor cost reductions No labor cost reductions Labor cost reductions No labor cost reductions
Actual Div. Below Expectation 11.59 23.91 16.05 30.25
(16) (33) (52) (98)
Actual Div. Equals Expectation 8.70 15.22 6.79 19.75
(12) (21) (22) (64)
Actual Div. Above Expectation 7.25 33.33 6.48 20.68
(10) (46) (21) (67)
Total 27.54 72.46 29.32 70.68
The table is based on the whole sample. The table categorizes the firms into two groups: firms with CEO conservatism > 0.5 and firms with CEO conservatism
≤ 0.5 during calendar year 2020. The main numbers are the numbers of firms with a particular combination of dividend and labor cost reduction decision
expressed as a percentage of the total number of CEOs with the level of CEO conservatism in question. For example, the top left cell reports the number of firms
that opted for labor cost reductions, while the firm’s actual dividend was below the expected one expressed as a percentage of all firms with CEO conservatism >
0.5. Because of rounding errors, the sum of all the percentages for a given type of CEO may not add up to 100%. The numbers in parentheses are the numbers
of firms that engage in specific combinations of (no) labor cost reductions and dividend decision.
Table 4
Ways of reducing labor costs by CEO conservatism.
Firms with CEO Conservatism > 0.5 Firms with CEO Conservatism ≤ 0.5
No Labor Cost Reductions 72.54 70.80
(103) (240)
Labor Cost Reductions 27.46 29.20
(39) (99)
Temporary 71.79 82.83
(28) (82)
Permanent 51.28 41.41
(20) (41)
Voluntary 28.21 17.17
(11) (17)
Compulsory 89.74 84.85
(35) (84)
Shortened Working Hours 10.26 20.20
(4) (20)
Reduced Salary 25.64 30.30
(10) (30)
Reduced Salary Board 53.85 44.45
(21) (45)
The table divides the sample firms into two subsamples: the subsample of firms with CEO conservatism > 0.5 and the subsample of firms with CEO conservatism
≤ 0.5 during calendar year 2020. It reports the percentages of firms with and without labor cost reductions for each subsample. In addition, it reports the
number of firms in each subsample that engaged in a particular way of reducing labor costs expressed as a percentage of the total number of subsample firms
that engaged in labor cost reductions. It should be noted that, because firms may engage in more than one way of reducing labor, the sum of the percentages
exceeds 100%. The numbers in parentheses in the table refer to the actual numbers of firms. There are some small differences in the percentages of firms that
do and do not reduce labor costs between this table and the previous table. These differences are due to missing data on the dividend decision.
Table 3 suggests that the percentage of firms with conservative CEOs that opt for ‘‘shareholder pain’’ – that is, firms that do not
meet dividend expectations while staying clear of downsizing – is only 24%, compared to about 30% for firms with nonconservative
CEOs. In turn, there is a greater percentage of firms with conservative CEOs (i.e., 15.95%; = 8.70% + 7.25%) opting for ‘‘employee
pain’’ – that is, firms that reduce labor costs while meeting or exceeding dividend expectations, compared to the equivalent
percentage (i.e., 13.27%; = 6.79% + 6.48%) of all other firms.
Table 4 reports the percentages of firms that do and do not reduce their labor costs, including the different ways of doing this,
while distinguishing between conservative CEOs (i.e., CEOs for which 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 exceeds 0.5) and all other CEOs. It should
be noted that as some firms engage in more than one way of reducing labor costs, the sum of the percentages of firms engaging
in the various ways of reducing labor costs exceeds 100%. While temporary labor cost reductions include mostly furloughing and
voluntary leaves from work, we also considered shortened working hours and salary reductions as temporary and compulsory labor
cost reductions.
The table suggests that firms with conservative CEOs are less likely to adopt temporary measures to reduce their labor costs and
more likely to adopt permanent measures when compared to the remaining firms. Of particular interest is the large difference in
the percentage of firms with conservative CEOs that reduce working hours (i.e., 10%, compared to 20% of the remaining firms).
4.3. Regression analysis
4.3.1. The reaction to the pandemic
Table 5 reports the results from the estimation of Eq. (1). In Panel A, the political ideology of the CEO is measured
by 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚; in Panel B, the CEO indicator variables are used (i.e., 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠, 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠, and 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠, with
12
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠 having been dropped to avoid perfect multicollinearity). It should be noted that the table reports the regression
coefficients rather than the marginal effects, which cannot be reported in a straightforward way for multinomial logits. Nevertheless,
Fig. E.1 in Appendix E shows the marginal effects, which we discuss below.
Panel A of Table 5 provides consistent evidence in favor of our main hypothesis. In detail, the first column suggests that greater
CEO conservatism increases the likelihood of a firm opting for ‘‘employee pain’’ (i.e., it reduces labor costs while meeting dividend
expectations) as compared with opting for ‘‘shareholder pain’’ (i.e., paying out a dividend that is below the expected dividend while
not engaging in labor cost reductions). Similarly, the second column suggests that CEO conservatism also increases the likelihood
of ‘‘employee pain’’ as compared to ‘‘shared pain’’. In other words, a more conservative CEO is more likely to pay a dividend that
meets the expected dividend while reducing labor costs compared to the alternative of failing to meet the expected dividend while
also reducing labor costs. Finally, the last column of Panel A suggests that firms with more conservative CEOs are more likely to
opt for ‘‘employee pain’’ than for ‘‘no pain’’, with the latter consisting of maintaining labor costs while paying out a dividend equal
to at least the expected dividend.27
Panel B of Table 5, which uses the indicator variables rather than the continuous index, confirms the results from Panel A.
First, the indicator variable for conservative CEOs is consistently positive and significant (at the 10% level or better) across all three
columns of the panel. This suggests that conservative CEOs were more likely to make their employees bear the negative consequences
of the pandemic when compared to all other CEOs. In detail, such CEOs were more likely to opt for ‘‘employee pain’’ rather than
the alternatives of ‘‘shareholder pain’’, ‘‘shared pain’’, or ‘‘no pain.’’ In contrast, neither of the other two indicator variables of CEO
political ideology – that is, 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠 or 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠 – is significant.
Importantly, in both panels, the indicator variable 𝐿𝑜𝑠𝑠 is positive and significant (at the 10% level or better), except in the middle
column of Panels A and B. This suggests that firms with negative EPS were more likely to reduce their labor costs. Interestingly,
the coefficient on the interaction between 𝐿𝑜𝑠𝑠 and 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 in Panel A is negative and significant in two regressions (at
the 10% level or better), while the coefficient on the interaction between 𝐿𝑜𝑠𝑠 and the indicator variable 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠 in Panel B
is negative and significant (at the 10% level or better) in all regressions. This suggests that conservative CEOs may have reduced
labor costs to avoid an earnings loss, which would then have enabled them to pay a dividend that met investor expectations from
the pre-pandemic year.
Finally, in both panels, we find some evidence that firms with greater 𝐶 𝑎𝑠ℎ ℎ𝑜𝑙𝑑 𝑖𝑛𝑔 𝑠 were less likely to let their employees bear
the pain of the pandemic (see the regression coefficient in the rightmost column).
Figs. E.1 (a) to E.1 (b) in Appendix E depict the marginal effects for the 95% confidence intervals for the 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠 indicator
variable on the predicted probability for (a) ‘‘shareholder pain’’ and (b) ‘‘employee pain’’ for the entire sample. Figs. E.1 (c) and
E.1 (d) show the equivalent effects for the firms that made an earnings loss in at least one of the quarters of calendar year 2020,
whereas Figs. E.1 (e) and E.1 (f) report the equivalent effects for the firms that did not make an earnings loss in any of the quarters
of calendar year 2020. The marginal effects confirm the results from Table 5.
In particular, Fig. E.1 (a) suggests that conservative CEOs have a 20% likelihood of inflicting shareholder pain compared to a
29% likelihood for nonconservative CEOs. In contrast, conservative CEOs have a 30% likelihood of reducing labor costs compared
to a corresponding likelihood of only 12% for nonconservative CEOs (see Fig. E.1 [b]). When focusing on the firms that made an
earnings loss in 2020, the likelihood of firms with a conservative CEO engaging in shareholder pain is virtually identical at 18%
(see Fig. E.1 [c]), whereas such firms are now more likely to engage in labor cost reductions. The likelihood of doing so increases
from 29% to 51% (see Fig. E.1 [d]). In turn, Fig. E.1 (e) suggests that firms that do not make a loss, independent of their CEO’s
political ideology, have a similar likelihood of engaging in shareholder pain when compared to loss-making firms. Finally, Fig. E.1
(f) suggests that in the absence of an earnings loss, conservative CEOs are more likely to reduce labor costs – the likelihood being
20% compared to a corresponding likelihood of only 7% for nonconservative CEOs. To sum up, the results in Table 5 are not only
statistically but also economically significant.
Both panels of Table 6 recode the reaction labeled ‘‘employee pain’’ by distinguishing between temporary and permanent
reductions in labor costs.28 As in the previous table, Panel A uses 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 to measure CEO political ideology, whereas
Panel B uses the CEO indicator variables (i.e., 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠, 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠, and 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠; again, 𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠 has been dropped to
avoid perfect multicollinearity). First, we investigate whether (more) conservative CEOs are more likely to use temporary reductions
in labor costs in order to meet dividend expectations as compared to maintaining labor costs while paying out a dividend per share
that falls below expectations (first column). Second, we investigate whether conservative CEOs are more likely to resort to permanent
rather than temporary reductions in labor costs when compared with all other CEOs (second column). For completeness, we also
check whether (more) conservative CEOs are more likely to use permanent reductions in labor costs in order to meet dividend
expectations as compared to no reductions in labor costs and not meeting dividend expectations (third column).
27 As discussed in Section 2, the CARES Act provided assistance, such as loans and tax support, for businesses affected by the COVID-19 pandemic. Importantly,
firms could not make involuntary dismissals or furloughs while receiving support under the CARES Act. Still, while the CARES Act typically limited compulsory
downsizing, Panel A of Table 5 suggests that the 𝐶 𝐴𝑅𝐸 𝑆 indicator variable increased the likelihood of the firm engaging in ‘‘employee pain’’ rather than
‘‘shareholder pain’’ at the 5% significance level. This counterintuitive result can be explained by some firms conducting downsizing in an earlier quarter and
then applying under the CARES Act and receiving support in later quarters. Other firms initially obtained support via the CARES Act, but because of the absence of
additional support, had to downsize. For example, in October 2020, American Airlines Group proceeded with furloughs to reduce its headcount absent an extension
of the CARES Act Payroll Support Program (PSP). See https://www.sec.gov/Archives/edgar/data/4515/000000620120000100/a8kerexhibit991q3-20.htm.
28 Note that the regressions in Table 6 contain the same control variables as the regressions in Table 5. However, these are not reported for the sake of
brevity. The full regression results can be found in Table 9 in the online appendix.
13
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 5
Likelihood of ‘‘employee pain’’ compared to ‘‘shareholder pain,’’ ‘‘shared pain,’’ and ‘‘no pain’’.
Panel A: Using CEO Conservatism
( ) ( ) ( )
𝑃𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑝𝑎𝑖𝑛 𝑃𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑝𝑎𝑖𝑛 𝑃𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑝𝑎𝑖𝑛
𝑙𝑛 𝑃𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑 𝑒𝑟 𝑝𝑎𝑖𝑛
𝑙𝑛 𝑃𝑠ℎ𝑎𝑟𝑒𝑑 𝑝𝑎𝑖𝑛
𝑙𝑛 𝑃𝑛𝑜 𝑝𝑎𝑖𝑛
Conservatism 1.904* 2.400** 2.156**
(0.050) (0.038) (0.045)
Loss*Conservatism −2.409 −1.187 −3.446**
(0.106) (0.462) (0.043)
Employee St. Dev. 0.022 −0.005 −0.003
(0.199) (0.743) (0.863)
Div. Sur. St. Dev. −0.256 −0.102 3.139***
(0.707) (0.897) (0.009)
EPS Surprise −0.068 −0.214 −0.309*
(0.690) (0.257) (0.097)
Loss 2.549** 0.712 2.918**
(0.014) (0.507) (0.013)
𝛥 EPS 0.057 −0.073 −0.022
(0.536) (0.519) (0.856)
EPS −0.129 −0.072 −0.139
(0.135) (0.508) (0.105)
Repurchases2020 0.888 0.485 0.682
(0.159) (0.475) (0.347)
Repurchases2019 −0.460 −0.983 −0.987
(0.504) (0.197) (0.225)
Productivity2019 −0.039 −0.099 −0.105
(0.849) (0.651) (0.630)
Inst. Ownership Ratio2019 0.295 0.231 0.379
(0.353) (0.497) (0.312)
CARES 1.715*** 1.161** 1.541**
(0.005) (0.045) (0.029)
Industry Labor Expense2019 −0.025 −0.022 −0.021
(0.360) (0.464) (0.489)
Red State −0.407 −0.575 −0.795*
(0.310) (0.185) (0.082)
CEO Age 0.001 −0.059* 0.001
(0.971) (0.078) (0.978)
CEO Gender 0.693 0.743 −0.955
(0.425) (0.410) (0.432)
CEO Duality 0.135 0.134 0.800*
(0.745) (0.768) (0.098)
MBA −0.000 −0.305 0.267
(1.000) (0.495) (0.572)
Ivy League 0.284 0.120 −0.315
(0.741) (0.897) (0.730)
Professional Qualification −0.735 −0.680 −1.170
(0.427) (0.482) (0.242)
CEO Share Ownership 0.692 −0.109 −0.388
(0.235) (0.827) (0.436)
Tobin’s Q −0.001 0.048 −0.075
(0.993) (0.780) (0.616)
Cash Holdings −2.326 −1.774 −3.755*
(0.289) (0.470) (0.098)
ESG Score −0.011* −0.009 −0.002
(0.083) (0.206) (0.786)
Credit Rating −0.251 0.346 −0.602
(0.688) (0.646) (0.397)
Firm Size2019 −0.289*** −0.278** −0.342***
(0.006) (0.031) (0.009)
Market to Book2019 −0.001 0.000 0.000
(0.884) (0.999) (0.928)
R&D2019 −0.000 0.000 0.000
(0.681) (0.272) (0.582)
(continued on next page)
14
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 5 (continued).
Firm Age2019 0.428* 0.062 0.547**
(0.071) (0.817) (0.032)
Foreign Sales2019 1.514* 2.579** 2.241**
(0.094) (0.017) (0.033)
Capital Expenditures2019 11.720* 7.955 2.737
(0.060) (0.232) (0.700)
Leverage2019 −0.037 −0.362 2.899**
(0.968) (0.725) (0.013)
Consumer 0.903 0.288 0.732
(0.114) (0.633) (0.277)
Manufacturing −0.362 −0.275 0.228
(0.525) (0.666) (0.752)
HiTec 0.933 0.197 −0.570
(0.179) (0.803) (0.434)
Health −0.037 −1.136 −1.866*
(0.972) (0.306) (0.091)
China 0.499 −0.135 0.165
(0.295) (0.799) (0.758)
Constant −3.142 4.305 −0.191
(0.248) (0.159) (0.952)
N 459 459 459
Pseudo R2 0.215 0.215 0.215
Likelihood Ratio 𝜒 2 309.543 309.543 309.543
Panel B: Using CEO Political Ideology Types
Liberals 0.569 0.663 0.216
(0.575) (0.627) (0.847)
Conservatives 1.708** 3.449*** 1.532*
(0.016) (0.006) (0.058)
Nonpartisans 0.739 0.565 −0.213
(0.294) (0.474) (0.778)
Loss*Liberals 13.993 −0.026 0.310
(0.983) (0.989) (0.881)
Loss*Conservatives −1.714 −3.345** −2.924**
(0.137) (0.036) (0.024)
Loss*Nonpartisans −0.154 −1.133 −1.157
(0.878) (0.262) (0.287)
Employee St. Dev. 0.022 −0.006 −0.004
(0.197) (0.664) (0.803)
Div. Sur. St. Dev. −0.314 −0.189 3.150***
(0.650) (0.814) (0.010)
EPS Surprise −0.070 −0.244 −0.334*
(0.685) (0.209) (0.079)
Loss 1.453** 0.790 1.731**
(0.018) (0.231) (0.019)
𝛥 EPS 0.077 −0.046 0.019
(0.405) (0.685) (0.884)
EPS −0.158* −0.101 −0.169*
(0.071) (0.364) (0.053)
Repurchases2020 0.996 0.684 0.839
(0.128) (0.331) (0.268)
Repurchases2019 −0.605 −1.301* −1.254
(0.395) (0.094) (0.139)
Productivity2019 −0.023 −0.075 −0.044
(0.910) (0.734) (0.840)
Inst. Ownership 0.302 0.238 0.372
Ratio2019
(0.336) (0.489) (0.329)
CARES 1.683*** 0.974 1.448**
(0.007) (0.100) (0.046)
Industry Labor −0.017 −0.006 −0.010
Expense2019
(0.548) (0.834) (0.740)
Red State −0.343 −0.530 −0.744
(0.395) (0.222) (0.105)
CEO Age 0.004 −0.064* 0.003
(0.893) (0.061) (0.923)
(continued on next page)
15
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 5 (continued).
CEO Gender 1.068 0.985 −0.982
(0.281) (0.336) (0.447)
CEO Duality 0.183 0.398 1.054**
(0.668) (0.402) (0.034)
MBA −0.039 −0.322 0.275
(0.926) (0.480) (0.567)
Ivy League 0.435 0.300 −0.156
(0.629) (0.755) (0.869)
Professional −0.662 −0.536 −0.980
Qualification
(0.481) (0.586) (0.336)
CEO Share Ownership 0.619 −0.197 −0.432
(0.287) (0.699) (0.396)
Tobin’s Q −0.000 0.044 −0.060
(0.998) (0.801) (0.692)
Cash Holdings −2.330 −2.085 −4.043*
(0.300) (0.405) (0.083)
ESG Score −0.011* −0.011 −0.004
(0.078) (0.115) (0.581)
Credit Rating −0.258 0.263 −0.757
(0.689) (0.735) (0.302)
Firm Size2019 −0.303*** −0.292** −0.352***
(0.004) (0.023) (0.007)
Market to Book2019 −0.001 −0.000 0.001
(0.895) (0.958) (0.862)
R&D2019 −0.000 0.000 0.000
(0.506) (0.250) (0.628)
Firm Age2019 0.446* 0.191 0.613**
(0.068) (0.481) (0.019)
Foreign Sales2019 1.327 2.197** 2.007*
(0.154) (0.043) (0.062)
Capital Expenditures2019 11.024* 7.551 2.665
(0.078) (0.263) (0.709)
Leverage2019 −0.133 −0.532 2.965**
(0.886) (0.620) (0.014)
Consumer 0.951 0.163 0.456
(0.106) (0.794) (0.516)
Manufacturing −0.382 −0.344 0.119
(0.514) (0.599) (0.871)
HiTec 0.978 0.285 −0.613
(0.164) (0.722) (0.406)
Health −0.239 −1.584 −1.918*
(0.824) (0.163) (0.087)
China 0.573 0.011 0.223
(0.245) (0.983) (0.687)
Constant −3.007 4.984 0.359
(0.267) (0.104) (0.909)
N 459 459 459
Pseudo R2 0.239 0.239 0.239
Likelihood Ratio 𝜒 2 343.873 343.873 343.873
The table reports the coefficients from estimating various multinomial logits explaining the likelihood of the firm adopting a (more) shareholder-friendly reaction
relative to a (more) employee-friendly reaction to the pandemic. Panel A uses CEO conservatism as the key variable of interest, whereas Panel B uses the indicator
variables for CEO political ideology (i.e., 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠, 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠, 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠, and 𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠). Employee pain refers to the firm reducing its labor costs while
paying out the expected dividend in each quarter of calendar year 2020. Shareholder pain refers to the firm not reducing its labor costs while paying out a
dividend that is below the 2020 dividend expected in 2019. Shared pain refers to the firm reducing its labor costs, while paying out a dividend that is below
the 2020 dividend expected in 2019. Finally, No pain refers to the firm not reducing its labor costs while paying out the expected dividend in each quarter of
calendar year 2020. The numbers in parentheses are the p-values. ∗, ∗∗, and ∗∗∗ refer to statistical significance at the 10%, 5%, and 1% levels, respectively.
Similar to the previous table, this table reports the regression coefficients rather than the marginal effects, which cannot be
reported in a straightforward way for multinomial logits. Nevertheless, Figs. E.2 (a) to E.2 (f) in Appendix E show the marginal
effects, which we discuss below.
The first column of Panel A of Table 6 supports the argument that (more) conservative CEOs use temporary reductions in labor
costs to avoid negative earnings per share, enabling them to meet dividend expectations. While the coefficient on 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚
is not significant, the coefficient on the indicator variable 𝐿𝑜𝑠𝑠 is positive and significant (at the 1% level), whereas the coefficient
on the interaction between 𝐿𝑜𝑠𝑠 and 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 is negative and significant (at the 10% level). This suggests that while firms
with nonconservative CEOs are more likely to use temporary reductions in labor costs when there is a loss, the opposite observation
applies to firms with conservative CEOs. Altogether, this confirms the argument that conservative CEOs temporarily reduce labor
costs to meet dividend expectations. Finally, the second column of Panel A of Table 6 provides evidence that conservative CEOs are
16
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 6
(
Likelihood of temporary and permanent reductions in labor costs.
) ( ) ( )
𝑃𝑡𝑒𝑚𝑝𝑜𝑟𝑎𝑟𝑦 𝑟𝑒𝑑 𝑢𝑐 𝑡𝑖𝑜𝑛𝑠 𝑃𝑝𝑒𝑟𝑚𝑎𝑛𝑒𝑛𝑡 𝑟𝑒𝑑 𝑢𝑐 𝑡𝑖𝑜𝑛𝑠 𝑃𝑝𝑒𝑟𝑚𝑎𝑛𝑒𝑛𝑡 𝑟𝑒𝑑 𝑢𝑐 𝑡𝑖𝑜𝑛𝑠
𝑙𝑛 𝑃𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑 𝑒𝑟 𝑝𝑎𝑖𝑛
𝑙𝑛 𝑃𝑡𝑒𝑚𝑝𝑜𝑟𝑎𝑟𝑦 𝑟𝑒𝑑 𝑢𝑐 𝑡𝑖𝑜𝑛𝑠
𝑙𝑛 𝑃𝑠ℎ𝑎𝑟𝑒ℎ𝑜𝑙𝑑 𝑒𝑟 𝑝𝑎𝑖𝑛
Panel A: Using CEO Conservatism
Conservatism 1.334 3.250* 2.868
(0.222) (0.083) (0.106)
Loss*Conservatism −3.278* −0.622 −2.235
(0.062) (0.796) (0.339)
Loss 3.017*** 0.761 2.865*
(0.008) (0.661) (0.096)
N 459 459 459
Pseudo R2 0.211 0.211 0.211
Likelihood Ratio 𝜒 2 321.744 321.744 321.744
Panel B: Using CEO Political Ideology Types
Liberals −0.028 2.209 1.991
(0.982) (0.249) (0.238)
Conservatives 1.020 4.890*** 3.395**
(0.194) (0.007) (0.024)
Nonpartisans −0.689 2.327* 2.458*
(0.466) (0.098) (0.071)
Loss*Liberals 15.181 −0.159 16.098
(0.993) (0.951) (0.993)
Loss*Conservatives −15.949 −3.005 −1.920
(0.982) (0.172) (0.315)
Loss*Nonpartisans −0.004 −1.198 −0.279
(0.998) (0.479) (0.869)
Loss 1.563** 1.285 2.065
(0.015) (0.373) (0.147)
N 459 459 459
Pseudo R2 0.238 0.238 0.238
Likelihood Ratio 𝜒 2 362.695 362.695 362.695
This table reports the coefficients from estimating two multinomial logits with two different base cases. The first column estimates the likelihood of the firm
reducing its labor costs temporarily to meet the expected dividend (Temporary reductions) compared to avoiding reductions in labor costs while paying out a
dividend below the expected dividend (Shareholder pain). The second column estimates the likelihood of permanent reductions in labor costs while meeting
the expected dividend (Permanent reductions) compared to Temporary reductions. The last column estimates the likelihood of Permanent reductions compared to
Shareholder pain. Panel A uses CEO conservatism as the key variable of interest, whereas Panel B uses the indicator variables for CEO political ideology (i.e.,
𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠, 𝐿𝑖𝑏𝑒𝑟𝑎𝑙𝑠, 𝑁 𝑜𝑛𝑝𝑎𝑟𝑡𝑖𝑠𝑎𝑛𝑠, and 𝑍 𝑒𝑟𝑜𝑑 𝑜𝑛𝑎𝑡𝑖𝑜𝑛𝑠). Temporary reductions refers to the firm reducing its labor costs temporarily (e.g., via furloughing) in
at least one of the four quarters of the 2020 calendar year, whereas Permanent reductions refers to the firm reducing its labor costs permanently in at least one
of the four quarters of the 2020 calendar year. Shareholder pain refers to the firm not reducing its labor costs while paying out a dividend that is below the
2020 dividend as expected in 2019 in at least one of the 2020 quarters. The regressions include the same control variables as those in Table 5, but these are
not reported for the sake of brevity. The numbers in parentheses are the p-values. ∗, ∗∗, and ∗∗∗ refer to statistical significance at the 10%, 5%, and 1% levels,
respectively.
more likely to use permanent rather than temporary reductions in labor costs. Albeit the coefficient in question just falls short of
significance at the 10% level, there is some suggestion in the last column that conservative CEOs are more likely to use permanent
reductions in labor costs than cause ‘‘shareholder pain’’.
Panel B of Table 6 confirms the results from Panel A. The indicator variable for conservative CEOs is positive and significant (at
the 5% level or better) across the last two columns of the panel. We find that conservative CEOs are more likely to use permanent
rather than temporary reductions in labor costs. In addition, conservative CEOs are more likely to permanently reduce their labor
costs than let their shareholders bear the cost of the pandemic.
Figs. E.2 (a) to E.2 (f) in Appendix E show the marginal effects for the 95% confidence intervals for the 𝐶 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑣𝑒𝑠 indicator
variable on the predicted probability for temporary and permanent reductions in labor costs for the entire sample (Figs. E.2 [a] and
E.2 [b]), for the subsample of firms that made an earnings loss in at least one of the quarters of calendar year 2020 (Figs. E.2 [c]
and E.2 [d]), and the subsample of firms that did not make an earnings loss in any of the quarters of calendar year 2020 (Figs. E.2
[e] and E.2 [f]). While the difference between the likelihood of conservative CEOs to engage in temporary labor reductions and
the corresponding likelihood for nonconservative CEOs is relatively small (see Fig. E.2 [a], where the difference is four percentage
points), the equivalent difference in the likelihoods for permanent reductions in labor costs is economically significant, as it amounts
to 18 percentage points (see Fig. E.2 [b]). When the subsample of firms that made an earnings loss in 2020 is considered (see Fig. E.2
[d]), this difference is even greater at 27 percentage points. Still, even in the absence of an earnings loss, conservative CEOs are
more likely to engage in permanent reductions in labor costs (see Fig. E.2 [f]), the difference in the likelihood being 12 percentage
points.
17
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
4.4. Robustness tests
We perform a battery of robustness tests. First, we reestimate the regressions from both Tables 5 and 6 by dropping the 48 firms
that applied for government support under the CARES Act. We find qualitatively similar results (not tabulated), albeit with the
interaction between 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 and 𝐿𝑜𝑠𝑠 being less significant29 and one regression (i.e., the second regression in Table 6)
struggling with high standard errors because of the substantial drop in observations.
Second, we include the pre-pandemic cumulative abnormal returns (CARs) and their equivalent for the pandemic period in the
regression analysis. The pre-pandemic period covers January 2, 2017, to December 31, 2019. Like Dechow et al. (2021), we define
the pandemic period as January 2, 2020, to March 31, 2020. The CARs are based on the Fama–French three-factor model (Fama
and French, 1993), and alternatively, the Fama–French–Carhart four-factor model (Carhart, 1997).30 When we include the CARs,
the number of observations in the regression analysis in Tables 5 and 6 drops from 459 to 418. These augmented regressions, which
are not tabulated, confirm our key results. As expected, the coefficients on the pre-pandemic and pandemic CARs – when significant
– are negative, suggesting that firms that experienced negative CARs in either or both periods were more likely to reduce their labor
costs during the calendar year 2020. Again, it should be noted that we find some evidence (at the 10% level) that conservative CEOs
are more likely to opt for permanent rather than temporary reductions in labor costs. Importantly, there is also evidence (again at
the 10% level) that conservative CEOs may engage in permanent downsizing to meet or exceed dividend expectations.
Third, 20 of the 459 sample firms operate in Investment Banking and Securities Dealing and Commercial Banking. Since the U.S.
government imposed caps on dividends and suspended stock repurchases for bank holding companies during the pandemic (Federal
Reserve, 2020), we rerun the regressions from Tables 5 and 6 by excluding these 20 firms from our sample (these regressions are
not tabulated). We find qualitatively similar results to those in our main analysis.
Fourth, we recode 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 by setting to 0.5 values of the index between 0.4 and 0.6. The rationale behind this is that
CEOs with values close to 0.5 do not have strong political leanings and may therefore be de facto nonpartisan. Put differently, we
now allow for a margin of error. We classify CEOs who made identical or nearly identical political contributions to both parties as
nonpartisan CEOs. When we reestimate the multinomial logits in Panel A of Table 5 (the results are not tabulated for the sake of
brevity), we still find results that are qualitatively similar to our main results.
Fifth, we reestimate the multinomial logits in Tables 5 and 6 by including the contemporaneous levels (i.e., the 2020 levels
of R&D expenditures, cash holdings, cash flow, and capital expenditures). The regression results (not tabulated) uphold our key
results regarding the effect of CEO political ideology. When it comes to the interaction(s) between CEO conservatism (the CEO
political ideology indicator variables) and the Loss indicator variable, we find that the coefficient(s) on the interaction(s) is (are) no
longer significant in some regressions, although they gain significance in one regression. However, this change in the results can be
explained by the high correlation (−0.488) between the loss indicator variable and cash flow in 2020.
Sixth, given that Compustat does not report labor costs for many firms, we use the logarithm of the industry average for the sum
of salaries and wages plus additional costs for such firms. However, when the 2019 labor costs are excluded from the regressions
in Tables 5 and 6, the results (which are not tabulated) do not change qualitatively.
Finally, it could be the case that CEO share ownership affects the likelihood that dividends met or exceeded expectations during
the pandemic. While the regressions in Tables 5 and 6 already include CEO share ownership, the coefficient on this variable was
consistently nonsignificant. Nevertheless, it could be the case that CEO share ownership becomes significant when this variable is
interacted with the loss indicator variable. While our original results are upheld, the coefficient on the interaction between CEO
share ownership and the loss indicator variable is not significant in any of these regressions (not tabulated).
4.5. Further analysis
Below, we perform additional analysis to investigate the following: (i) whether CEO political ideology had an effect on alternative
outcome variables (i.e., the number of employees, dividend surprise, and restructuring costs, all relating to the year 2020), (ii)
whether CEO political ideology affected the likelihood of labor cost reductions combined with a salary cut for the board when
compared to (a) combined labor cost reductions without a salary cut for the board and (b) no labor cost reductions, and (iii)
whether the various categories of CEO and types of reactions affected firm value and firm performance in 2021 and 2022.
The rationale behind using alternative outcome variables is twofold. First, this exercise enables us to assess the economic
significance of CEO political ideology, which cannot be determined via the use of our main dependent variable. Second, when
it comes to the effect of CEO political ideology on the number of employees, it also enables us to evaluate value added from our
granular dataset. Indeed, the main analysis suggests that S&P 500 firms used various approaches to reduce their labor costs, some of
which did not affect employee numbers (e.g., reductions in the number of hours worked). Hence, the question arises as to whether
CEO political ideology has similar explanatory power for the number of employees or whether the explanatory power is limited to
our measure of broader labor cost reductions, which include reductions that do not necessarily affect the number of employees.
Table 7 reports the ordinary least squares (OLS) regressions explaining the natural logarithm of the number of employees in
2020 in the first column, the natural logarithm of one plus the dividend surprise in 2020 in the second column, and the natural
29 This result can be easily explained by the significant and positive correlation between the 𝐶 𝐴𝑅𝐸 𝑆 indicator variable and the 𝐿𝑜𝑠𝑠 indicator variable. See
Table 2.
30 Further details about the estimation window, and so on, can be found in Appendix D.
18
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 7
Regressions estimating number of employees, restructuring costs, and dividend surprise in 2020.
Number of Restructuring Dividend
Employees2020 Costs2020 Surprise2020
Conservatism 0.008 0.776** 0.040**
(0.838) (0.043) (0.019)
Loss*Conservatism −0.064 −0.509 −0.036
(0.376) (0.450) (0.223)
Number of Employees2019 0.973***
(0.000)
Restructuring Costs2019 0.569***
(0.000)
Dividend Surprise2019 0.682***
(0.000)
EPS Surprise 0.005 −0.054 −0.001
(0.406) (0.360) (0.774)
Loss −0.056 1.141** 0.009
(0.246) (0.013) (0.648)
N 442 459 429
R2 0.986 0.558 0.542
This table reports the coefficients of three regressions explaining the natural logarithm of the number of employees in 2020 in
the first column, the natural logarithm of one plus the 2020 restructuring costs in the second column, and the natural logarithm
of one plus the dividend surprise in 2020 in the last column. CEO conservatism is the key variable of interest. The regressions
include the same control variables as those in Table 5, but these are not reported for the sake of brevity. The numbers in
parentheses are the p-values. ∗, ∗∗, and ∗∗∗ refer to statistical significance at the 10%, 5%, and 1% levels, respectively.
logarithm of one plus the 2020 restructuring costs in the final column.31 As the first column suggests, there is no significant effect
of CEO conservatism on the number of employees. This highlights the importance of the granularity of our data, especially in light
of the temporary nature of the pandemic, as well as the much more sudden, large economic fallout compared to, for example,
the Great Recession. In contrast, we find a significant and positive effect of CEO conservatism on the restructuring costs and the
dividend surprise. The regression results suggest that an increase of one standard deviation in the CEO conservatism index (the
standard deviation is 0.273, per Table 1) results in an increase in the 2020 restructuring costs of 22.8%. As the sample average for
the 2020 restructuring costs is US$ 134.7 million, this corresponds to an increase of US$ 30.7 million. Hence, the effect of CEO
conservatism is also economically significant. When it comes to the economic significance of the regression results for the dividend
surprise, the third column of Table 7 shows that an increase of one standard deviation in the CEO conservatism index (again, the
standard deviation is 0.273) results in a 1.1% increase in the dividend surprise.
Note that we do not find that CEO political ideology affects the likelihood of labor cost reductions being combined with a salary
cut for the board when compared to (i) labor cost reductions without a salary cut for the board and (ii) no labor cost reductions.
These results are not tabulated.
Finally, we investigate whether the various categories of CEO, as well as the main types of reaction to the pandemic
(i.e., employee vs. shareholder pain) affected firm value and performance following the pandemic (i.e., in the fiscal years 2021
and 2022). To this end, we perform an additional regression analysis, which is reported in Table 8. The regressions in Panel A use
𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚, whereas the regressions in Panel B use the CEO political ideology indicator variables. The regressions in the first
two columns of each panel explain firm value, measured by Tobin’s Q in 2022 and 2021 against firm performance in 2021 and
2020, respectively, and two indicator variables stating whether the firm reduced its labor costs in 2020 (𝑅𝑒𝑑 𝑢𝑐 𝑒𝑑 𝐿𝑎𝑏𝑜𝑟 𝐶 𝑜𝑠𝑡𝑠) and
whether the firm did not meet dividend expectations in the same year (𝐴𝑐 𝑡𝑢𝑎𝑙 𝐷𝑖𝑣. 𝐵 𝑒𝑙𝑜𝑤 𝐸 𝑥𝑝.). The regressions also include the
interaction of these two indicator variables. The equivalent regressions for the return on assets (ROA) can be found in the third and
fourth columns, while the equivalent regressions for the return on equity (ROE) are reported in the final two columns of Table 8.
Both variables are defined in Appendix D.
There is evidence in Panel A using 𝐶 𝐸 𝑂 𝑐 𝑜𝑛𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑠𝑚 that firms with conservative CEOs had a higher Tobin’s Q (the coefficient
is just not significant at the 10% level) and a significantly higher (at the 5% level) ROA in 2021. Note that the coefficient on the
indicator variable for reduced labor costs is also significant (at the 5% level) for the same regression explaining ROA in 2021. In
addition, in Panel B, firms with conservative CEOs have a significantly higher Tobin’s Q (at the 10% level) and a higher ROA (the
coefficient is just not significant at the 10% level) in 2021. Again, the coefficient on the indicator variable for reduced labor costs is
also significant (at the 5% level) for the same regression explaining ROA in 2021. All in all, these results suggest that conservative
CEOs did a somewhat better job when it came to the financial performance of their firms for 2021.
31 Note that the regressions in Table 7 contain the same control variables as the regressions in Table 5. However, these are not reported for the sake of
brevity. The full regression results can be found in Table 10 in the online appendix.
19
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 8
Regressions explaining firm value and firm performance by CEO political ideology.
Panel A: Using CEO conservatism
Tobin’s Q ROA ROE
2022 2021 2022 2021 2022 2021
Conservatism 0.132 0.295 0.002 0.018** −0.002 0.017
(0.389) (0.101) (0.878) (0.045) (0.871) (0.229)
Tobin’s Q2021 0.694***
(0.000)
Tobin’s Q2020 1.127***
(0.000)
ROA2021 0.761***
(0.000)
ROA2020 0.541***
(0.000)
ROE2021 0.383***
(0.000)
ROE2020 0.128***
(0.000)
Reduced Labor Costs −0.023 0.120 0.012 0.015** 0.008 −0.019
(0.859) (0.426) (0.181) (0.046) (0.488) (0.111)
Actual Div. Below Exp. 0.100 0.016 0.007 0.005 0.007 0.004
(0.323) (0.893) (0.301) (0.409) (0.406) (0.659)
Red.Lab.Cost*Act.Div.Below.Exp. 0.007 −0.177 −0.023* −0.008 −0.006 0.007
(0.969) (0.411) (0.076) (0.460) (0.686) (0.678)
Consumer 0.416*** −0.098 −0.002 0.043*** 0.007 −0.017
(0.001) (0.498) (0.869) (0.000) (0.472) (0.110)
Manufacturing 0.178 −0.081 0.022*** 0.039*** 0.022** 0.011
(0.105) (0.536) (0.004) (0.000) (0.018) (0.268)
HiTec 0.083 0.331** −0.009 0.042*** −0.015 −0.002
(0.512) (0.024) (0.325) (0.000) (0.154) (0.823)
Health 0.118 −0.216 −0.012 0.043*** −0.016 −0.002
(0.483) (0.277) (0.323) (0.000) (0.266) (0.902)
Constant 0.233* −0.243 0.007 0.002 0.028** 0.038***
(0.081) (0.135) (0.418) (0.813) (0.011) (0.001)
N 427 459 387 436 429 436
R2 0.789 0.804 0.509 0.531 0.184 0.082
Panel B: Using CEO Political Ideology Types
Tobin’s Q ROA ROE
2022 2021 2022 2021 2022 2021
Liberals 0.105 −0.146 0.018 −0.010 0.013 −0.020
(0.571) (0.481) (0.176) (0.317) (0.429) (0.217)
Conservatives 0.047 0.268* 0.010 0.012 0.003 0.009
(0.702) (0.066) (0.245) (0.102) (0.774) (0.401)
Nonpartisans 0.046 0.174 0.005 −0.005 −0.006 −0.012
(0.654) (0.150) (0.456) (0.436) (0.497) (0.195)
Tobin’s Q2021 0.695***
(0.000)
Tobin’s Q2020 1.130***
(0.000)
ROA2021 0.765***
(0.000)
ROA2020 0.542***
(0.000)
ROE2021 0.381***
(0.000)
ROE2020 0.130***
(0.000)
Reduced Labor Costs −0.024 0.118 0.012 0.016** 0.008 −0.018
(0.853) (0.436) (0.169) (0.041) (0.474) (0.128)
Actual Div. Below Exp. 0.105 0.027 0.009 0.005 0.008 0.004
(0.306) (0.822) (0.216) (0.417) (0.366) (0.703)
Red.Lab.Cost*Act.Div.Below.Exp. −0.001 −0.186 −0.023* −0.008 −0.006 0.007
(0.997) (0.387) (0.066) (0.471) (0.701) (0.657)
Consumer 0.422*** −0.061 −0.001 0.042*** 0.006 −0.020*
(0.001) (0.677) (0.943) (0.000) (0.573) (0.069)
(continued on next page)
20
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table 8 (continued).
Manufacturing 0.179 −0.078 0.022*** 0.038*** 0.021** 0.010
(0.105) (0.553) (0.004) (0.000) (0.024) (0.315)
HiTec 0.068 0.335** −0.010 0.041*** −0.016 −0.003
(0.594) (0.022) (0.272) (0.000) (0.130) (0.782)
Health 0.111 −0.253 −0.013 0.044*** −0.015 0.000
(0.514) (0.205) (0.286) (0.000) (0.302) (0.997)
Constant 0.286*** −0.163 0.004 0.012** 0.027*** 0.051***
(0.008) (0.213) (0.581) (0.030) (0.002) (0.000)
N 427 459 387 436 429 436
R2 0.789 0.805 0.513 0.533 0.187 0.089
This table reports the coefficients for regressions explaining Tobin’s Q, ROA, and ROE as a function of CEO conservatism in the years 2022 and 2021. The
numbers in parentheses are the p-values. ∗, ∗∗, and ∗∗∗ refer to statistical significance at the 10%, 5%, and 1% levels, respectively.
5. Discussion and conclusion
In this paper, we studied how the CEOs of S&P 500 firms reacted to the COVID-19 pandemic, distinguishing between (more)
shareholder-friendly reactions and (more) employee-friendly reactions. Shareholder-friendly reactions are those that prioritize
meeting dividend expectations over safeguarding jobs in the firm. In contrast, employee-friendly reactions prioritize safeguarding
jobs. We argued that during the pandemic, the CEO’s political ideology affected the choice between shareholder-friendly and
employee-friendly reactions. We hypothesized that conservative CEOs were more likely to prioritize the interests of the shareholders
over those of the workers during the height of the pandemic (i.e., during the calendar year 2020).
We used political donations made by CEOs during their lifetime up to and including the calendar year 2019 as a measure of their
political ideology. Our study benefits from granular data on labor cost reductions obtained from Form 8-K filings and quarterly data
on actual dividends per share and dividends per share forecasts for 2020 made at the end of 2019. Our dataset covers a range of
different forms of labor cost reductions, including those that affected worker numbers (e.g., permanent downsizing) and those that
did not (e.g., reductions in the number of working hours per week). We uncovered several findings. First, we consistently found that
conservative CEOs favored shareholders over workers during the pandemic. Second, when deciding on quarterly dividends for 2020,
CEOs used the consensus dividend per share for the various quarters of 2020 (as forecast by financial analysts in 2019) rather than
the actual quarterly dividends per share as their target dividends. Importantly, ceteris paribus, conservative CEOs were more likely
to pay out a dividend per share equal to the expected dividend per share with the workforce facing possible labor cost reductions,
including downsizing. Finally, we also found some evidence that conservative CEOs used temporary downsizing to reduce labor
costs, thereby avoiding negative earnings per share; this, in turn, enabled them to meet dividend expectations.
Our paper makes an important contribution to the sparse literature on how extreme events affect corporate decision-making by
studying how CEOs with different political ideologies reacted to the COVID-19 pandemic. Did they let investors bear the brunt of
the pandemic, or did they allow workers to do so?
Our paper also contributes to the literature on dividend and payout policy. Again, our findings showed that CEOs used the
dividend forecasts for 2020 made in the pre-pandemic year as their benchmark to decide on the 2020 dividends. This contrasts
with Lintner (1956), who argues that the current dividend level is guided by past dividend levels. Furthermore, our paper makes an
important contribution to the growing literature suggesting that CEO characteristics, including CEO political orientation, affect firm
strategy and decision-making. In contrast to previous research describing differences in behavior between conservative and liberal
CEOs, our findings suggest that conservative CEOs as a category are different from all other categories of CEOs, not just liberal ones,
in their reaction to an extreme event. Finally, our paper suggests that labor cost reductions and dividend decisions during periods
of crisis should not be studied in isolation, as the findings illustrate that CEOs considered both options during the crisis.
CRediT authorship contribution statement
Ali Bayat: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation,
Conceptualization. Marc Goergen: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project
administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Panagiotis
Koutroumpis: Writing – review & editing, Writing – original draft, Visualization, Methodology, Formal analysis, Data curation.
Xingjie Wei: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data curation.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared
to influence the work reported in this paper.
21
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Acknowledgments
We would like to express our heartfelt thanks to Tracy Wang, the Editor, and an anonymous referee for the helpful guidance we
received. We are grateful for helpful comments and suggestions to participants at the 7th annual conference of the International
Corporate Governance Society at the University of Groningen, the 12th Financial Markets and Corporate Governance Conference
at Monash University, the 2022 Finance Forum, the 2022 British Academy of Management Conference, the 22nd Hellenic Finance
and Accounting Association conference, and the 2023 World-Class Workshop in Finance at Edinburgh University Business School.
We would also like to thank seminar participants at the universities of Aberdeen, Leeds, Sabanci, and Sussex, as well as participants
at the 2022 Workshop on Corporate Governance and Investment at Copenhagen Business School, for their comments on earlier
versions of this paper. Special thanks are due to Miguel García-Cestona, Ricard Gil, Jordan Neyland, and Hanwen Sun. The research
reported in this paper was partially funded by the Spanish State Research Agency (AEI)/10.13039/501100011033) – Grant No.
PID2020-115982RB-C22 and Grant No. PID2023-148086NB-C22 granted by MCIU/AEI/10.13039/501100011033/FEDER, UE.
Appendix A. Sample selection process
The steps in the sample selection of the S&P firms for the period of the COVID-19 pandemic only (i.e., calendar year 2020) and
the resultant number of firm-year observations lost are as follows:
1. We excluded 19 firms whose headquarters are not located in the U.S. The remaining number of S&P 500 firms was 481.
We downloaded the Form 8-Ks for each of these firms during the four quarters of calendar year 2020 and coded their downsizing
variables.
2. Twenty-two firms were dropped after merging the data from Execucomp, I/B/E/S, and the downsizing data with the data
from Compustat. The final sample included 459 firms.
Appendix B. Excerpts from 8-K forms and mapping of labor cost reduction variables
See Table B.1.
Appendix C. Multinomial logit
The first multinomial logit uses the absence of labor cost reductions and the actual dividend being equivalent to the expected
dividend as the base case. In particular, the values that the dependent variable of this multinomial logit can take range from 0 to
4, and their corresponding reaction type is defined in the following table (see Table C.1):
Appendix D. The definition of variables
Dependent Variables
See Table C.1 in Appendix C for the definition of our main dependent variable.
Dividend measures
• Dividend exceeds expectations: An indicator variable that is set to one if the DPS for at least one of the four quarters of
the 2020 calendar year is higher than the expected DPS for that quarter (the expected DPS being the latest available in the
calendar year 2019) and zero otherwise (Source: Compustat and I/B/E/S).
• Dividend meets expectations: An indicator variable that is set to one if the DPS for at least one of the four quarters of the
2020 calendar year is equal to the expected DPS for that quarter (the expected DPS being the latest available in the calendar
year 2019) and zero otherwise (Source: Compustat and I/B/E/S).
• Dividend below expectations: An indicator variable that is set to one if the DPS for at least one of the four quarters of the
2020 calendar year is lower than the expected DPS for that quarter (the expected DPS being the latest available in the calendar
year 2019) and zero otherwise (Source: Compustat and I/B/E/S).
Measures for labor cost reductions
• Labor cost reductions: An indicator variable that is set to one if labor costs were reduced in any of the following ways in
at least one of the four quarters of the 2020 calendar year and zero otherwise. A reduction in salary is viewed as a type of
downsizing (Source: Form 8-K).
– Temporary: An indicator variable that is set to one if there was a temporary reduction in labor costs (e.g., via
furloughing) in at least one of the four quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
– Permanent: An indicator variable that is set to one if there was a permanent reduction in labor costs in at least one of
the four quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
– Voluntary: An indicator variable that is set to one if employees were offered voluntary leave in at least one of the four
quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
22
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Table B.1
Excerpts from 8-K forms and mapping of labor cost reduction variables.
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
1 0 1 1 1 0 0 0
Excerpt The company’s structural changes will result in the reallocation of some people and resources, which will include voluntary and
involuntary reductions in employees ... The voluntary program will first be offered to approximately 4,000 employees in the United
States, Canada and Puerto Rico who have a most-recent hire date on or before September 1, 2017.
Source The Coca-Cola Company, https://www.sec.gov/ix?doc=/Archives/edgar/data/21344/000155278120000463/e20478_ko-8k.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
1 1 0 0 0 0 0 0
Excerpt During second quarter of 2020, there were on average 307 employees on temporary emergency reserve boards, who are included in the
average employee count for the nine months ending September 30, 2020 in the above table. Employees on these boards received 7 days of
guarantee pay in each 28-day cycle and were eligible, if qualified, for Railroad Retirement Board unemployment benefits for the other 21
days, during which they were considered furloughed. Excluding these employees, the average headcount for the nine months ending
September 30, 2020, would be 19,658.
Source CSX Corporation, https://www.sec.gov/ix?doc=/Archives/edgar/data/277948/000027794820000046/csx-20201021.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
1 1 1 0 1 0 0 1
Excerpt Cost containment efforts included furloughs, layoffs, elimination of non-essential travel and capital expenditures, and temporary salary
reductions for upper management...
Source Rollins, Inc., https://www.sec.gov/ix?doc=/Archives/edgar/data/84839/000117120020000486/i20444_rol-8k.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
1 0 0 0 0 1 0 0
Excerpt Due to fact that a significant majority of the Company’s stores were closed or operating under shortened operating hours over the course
of the second half of the fiscal year, net sales changes for the Company, and each segment is based on absolute sales dollar changes and is
not presented in accordance with the Company’s comparable sales definition utilized in prior quarters.
Source Tapestry, Inc., https://www.sec.gov/Archives/edgar/data/1116132/000115752320001178/a52265820ex99_1.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
1 0 1 0 0 0 1 0
Excerpt . . . we had strong operating leverage on our expense for base payroll, benefits and tax at 150 basis points, which reflects decreased
headcount and temporary salary reductions across a broad range of our agencies and employees in response to the pandemic.
Source The Interpublic Group of Companies, Inc.,
https://www.sec.gov/ix?doc=/Archives/edgar/data/51644/000005164420000055/ipg-20201021.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
0 0 0 0 0 0 0 0
Excerpt Headcount increased 5% compared with the end of the second quarter of 2020, primarily reflecting the timing of campus hires.
Source The Goldman Sachs Group, Inc., https://www.sec.gov/Archives/edgar/data/886982/000119312520268983/d38229dex991.htm
Indicator Labor cost Temporary Permanent Voluntary Compulsory Shortened Reduced salary Reduced salary
variables reduction board
0 0 0 0 0 0 0 0
Excerpt Operating profit decreased 7% to $151 million and the operating profit margin declined 510 basis points to 64.6% primarily due to
increased legal related costs, higher compensation from increased headcount and incentives, as well as professional fees.
Source S&P Global Inc., https://www.sec.gov/ix?doc=/Archives/edgar/data/64040/000006404020000180/spgi-20201027.htm
Table C.1
Definition of the main dependent variable.
Dependent variable (value) Reaction label Reaction definition
0 No pain The firm does not reduce its labor costs in 2020 and pays out the 2020
dividend as expected in 2019 in each quarter of calendar year 2020.
1 Shareholder pain The firm does not reduce its labor costs, and its actual dividend is
below the expected dividend in at least one of the quarters of 2020.
2 No employee pain and The firm does not reduce its labor costs, and its actual dividend is
shareholder joy above the expected dividend and/or it repurchases shares (while its
actual dividend is not below the expected dividend) in at least one of
the quarters of 2020.
3 Employee pain The firm reduces its labor costs, but it pays out a dividend equal to or
greater than the expected dividend in each quarter of 2020.
4 Shared pain The firm reduces its labor costs, and its actual dividend is below the
expected dividend in at least one of the quarters of 2020.
We estimate various multinomial logits varying according to their base case — that is base case 0, base case 1, and base case 4.
23
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
– Compulsory: An indicator variable that is set to one if there were compulsory redundancies in at least one of the four
quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K and Compustat).
– Shortened: An indicator variable that is set to one if the firm applied a shortened workweek in at least one of the four
quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
– Reduced salary: An indicator variable that is set to one if the firm reduced salaries for all staff in at least one of the
four quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
– Reduced salary board: An indicator variable that is set to one if the firm reduced the salaries for the board only in at
least one of the four quarters of the 2020 calendar year and zero otherwise (Source: Form 8-K).
Key Variables
• CEO conservatism: Considering all the donations made by a CEO during their entire lifetime up to and including calendar
year 2020, but excluding the donations made one year before the U.S. presidential elections, CEO conservatism is measured as
the total amount donated to the Conservative party divided by the sum of the total amounts donated to the Republican Party
and Democratic Party. For CEOs who did not make any donations, this variable is set to 0.5 (Source: FEC).
• CEO political ideology — CEO types: A set of four indicator variables based on the following types of CEOs:
– Conservatives: An indicator variable taking the value of one if CEOs’ donations were to the Republican Party only and
zero otherwise (Source: Federal Election Commission [FEC]).
– Liberals: An indicator variable taking the value of one if CEOs’ donations were to the Democratic Party only and zero
otherwise (Source: FEC).
– Nonpartisans: An indicator variable taking the value of one if CEOs’ donations were to both the Democratic and
Republican parties and zero otherwise (Source: FEC).
– Zerodonations: An indicator variable taking the value of one if CEOs made no donations to any political party and zero
otherwise (Source: FEC).
Control Variables
• ∆EPS: EPS2020 – EPS2019 (Source: Compustat and the researcher’s own calculations).
• Capital expenditures: The ratio of capital expenditures over total assets for the fiscal year 2019 (Source: Compustat).
• CARES: An indicator variable that is set to one if the firm applied for assistance under the CARES Act during calendar year
2020 and zero otherwise (Source: Form 8-K and the U.S. Department of the Treasury website).
• Cash flow: Cash flow is defined as net cash flow from operating activities less cash flow from extraordinary items and
discontinued operations (Source: Compustat).
• Cash holdings: The ratio of cash and short term investments over total assets for the fiscal year 2019 (Source: Compustat).
• CEO age: The age of the CEO in 2019 (Source: Execucomp and BoardEx).
• CEO duality: An indicator variable that is set to one if the CEO is also the chair and zero otherwise (Source: BoardEx).
• CEO gender: An indicator variable that is set to one if the CEO is male and zero if female (Source: Form 8-K and the U.S.
Department of the Treasury website).
• CEO share ownership: The natural logarithm of one plus the percentage of shares outstanding held by the CEO at year end
2019. We set the value of the percentage of shares outstanding held by the CEO to zero for 18 observations with missing
values (Source: Execucomp).
• Credit rating: An indicator variable that is set to one if firms have a credit rating of investment grade for 2019 (i.e., a grade
rating of at least BBB-) and zero otherwise (Source: Capital IQ).
• Div. Sur. St. Dev.: The standard deviation of the dividend surprise for 2015–2019. The standard deviation is computed for
firms with at least three observations for the percentage change in the number of employees or the dividend surprise for the
period 2015–2019, and it is set to zero otherwise (Source: Compustat).
• Employee St. Dev.: The standard deviation of the percentage change in the number of employees for 2015–2019. The standard
deviation is computed for firms with at least three observations for the percentage change in the number of employees or the
dividend surprise for the period 2015–2019, and it is set to zero otherwise (Source: Compustat).
• EPS surprise: The difference between EPS for fiscal year 2020 and the expected EPS for the same year (Source: Compustat,
I/B/E/S, and the researcher’s own calculations).
• EPS: Earnings per share (EPS) for fiscal year 2020 (Source: Compustat).
• ESG score: This score ranges from 0 to 100, with scores between zero and 25 indicating poor ESG performance and insufficient
transparency in reporting ESG data, scores above 25 and up to 50 indicating satisfactory ESG performance and moderate
transparency in reporting ESG data, scores above 50 and up to 75 indicating good ESG performance and above-average
transparency in reporting ESG data, and scores above 75 indicating excellent ESG performance and high transparency in
reporting ESG data (Source: ASSET4/Refinitiv ESG).
• Firm age: The natural logarithm of the number of years that the firm has been covered by CRSP up to and including the year
2019 (Source: Compustat).
• Firm size: The natural logarithm of total assets for the fiscal year 2019 (Source: Compustat).
24
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
• Foreign sales: The ratio of foreign sales over foreign and domestic sales for the fiscal year 2019 (Source: Compustat).
• Industry indicator variables: We assign each firm to an industry based on its four-digit Standard Industry Classification
(SIC) code in 2020 based on the Fama and French five industries. The industries are Consumer, Manufacturing, HiTec, Health,
and Other (Source: Compustat and Kenneth French’s data library at http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/
data_library.html).
• Industry labor expense2019 : The natural logarithm of wages in fiscal year 2019. If the data are missing in Compustat, we use
data from the Annual Survey of Manufactures (ASM) conducted by the U.S. Census Bureau. More specifically, we estimate the
total cost to the employer, using data for the U.S. at the two-digit NAICS level as the industry average of the sum of salaries
and wages plus additional costs. That is the sum of the annual payroll, total fringe benefits, the employer’s cost for health
insurance, the employer’s cost for defined benefit pension plans, and the employer’s cost for other fringe benefits (Source:
Compustat and U.S. Census Bureau).
• Institutional ownership ratio2019 : The ratio of institutional ownership at the end of calendar year 2019 to the total shares
outstanding at the end of the same calendar year (Source: Thomson Reuters Institutional (13f) Holdings).
• Ivy League:
An indicator variable that is set to one if the CEO has graduated from an Ivy League university and zero otherwise (Source:
BoardEx).
• Leverage: The book value of long-term debt over the book value of total assets for the fiscal year 2019 (Source: Compustat).
• Loss: An indicator variable that is set to one if EPS is negative in at least one of the four quarters of calendar year 2020 and
zero otherwise (Source: Compustat).
• Market to book: The ratio of total market value of equity over book value of equity for the fiscal year 2019 (Source:
Compustat).
• MBA: An indicator variable that is set to one if the CEO has an MBA degree and zero otherwise (Source: BoardEx).
• Pandemic CARs: The cumulative abnormal returns (CARs) based on the Fama–French three-factor model and the Fama–
French–Carhart four-factor model, respectively. The pandemic period is defined as in Dechow et al. (2021), and it runs from
January 2, 2020, to March 31, 2020. The estimation window has a length of 250 trading days, and it ends 50 trading days
before the start of the event window. A minimum of 70 non-missing return observations within the estimation window is
required (Source: CRSP).
• PhD: An indicator variable that is set to one if the CEO has a PhD degree and zero otherwise (Source: BoardEx).
• Pre-pandemic CARs: The cumulative abnormal returns (CARs) based on the Fama–French three-factor model (Fama and
French, 1993) and the Fama–French–Carhart four-factor model (Carhart, 1997), respectively. The pre-pandemic period covers
January 2, 2018, to December 31, 2019. The estimation window has a length of 250 trading days, and it ends 50 trading days
before start of the event window. A minimum of 70 non-missing return observations within the estimation window is required
(Source: CRSP).
• Productivity2019 : The natural logarithm of net sales turnover in calendar year 2019 (Source: Compustat).
• Professional qualification: An indicator variable that is set to one if the CEO has a chartered accountant or certified public
accountant qualification and zero otherwise (Source: BoardEx).
• R&D: Research and development expenses for the fiscal year 2019. This variable is set to zero if the firm does not report this
item. (Source: Compustat).
• Red state: An indicator variable that is set to one if the firm’s headquarters are located in a state where a majority voted for
Republicans in the 2019 elections and zero otherwise.
• Repurchases𝑡 : An indicator variable that is set to one if there were stock repurchases in calendar year 𝑡 (𝑡 = 2019, 2020) and
zero otherwise (Source: Compustat).
• ROA: The return on assets calculated as the ratio of earnings before interest and taxes (EBIT) to total assets.
• ROE: The return on equity calculated as the ratio of net income (loss) to common shares outstanding multiplied by the stock
close price.
• Shares repurchased𝑡 : The ratio of total shares repurchased in calendar year 𝑡 to common shares outstanding in year 𝑡 − 1
(Source: Compustat).
• Tobin’s Q: The ratio of total assets plus the product of common shares outstanding and the 2019 end-of-year closing price
minus the book value of common equity over total assets. All accounting variables refer to the fiscal year 2019. (Source:
Compustat).
• Value shares repurchased𝑡 : The product of total shares repurchased in calendar year 𝑡 and the average price paid for the
repurchased shares (Source: Compustat).
Appendix E. Marginal effects
See Figs. E.1 and E.2.
Appendix F. Supplementary data
Supplementary material related to this article can be found online at https://doi.org/10.1016/j.jcorpfin.2024.102692.
25
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Fig. E.1. Marginal effects for multinomial logit from Panel A of Table 5.
This set of figures reports the marginal effects for the 95% confidence intervals for the Conservatives indicator variable based on the multinomial logit in Panel
A of Table 5. In the figure, (a) and (b) show the marginal effect of Conservatives switching from a zero to a one based on the likelihood of Shareholder pain and
the likelihood of Employee pain, respectively, for the entire sample. Figs. E.1 (c) and (d) are the equivalents for the subsample of firms with an earnings loss in
at least one quarter during calendar year 2020, and Figs. E.1 (e) and (f) are the equivalents for the subsample of firms without an earnings loss during calendar
year 2020.
26
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
Fig. E.2. Marginal effects for multinomial logit from Panel A of Table 6.
This set of figures reports the marginal effects for the 95% confidence intervals for the Conservatives indicator variable based on the multinomial logit in Panel
A of Table 6. Figs. E.2 (a) and (b) show the marginal effect of Conservatives switching from a zero to a one based on the likelihood of Temporary reductions
and the likelihood of Permanent reductions in labor costs, respectively, for the entire sample. Figs. E.2 (c) and (d) are the equivalent figures for the subsample
of firms with an earnings loss in at least one quarter during calendar year 2020, and Figs. E.2 (e) and (f) are the equivalent figures for the subsample of firms
without an earnings loss during calendar year 2020.
Data availability
The authors are unable or have chosen not to specify which data has been used.
27
A. Bayat et al. Journal of Corporate Finance 90 (2025) 102692
References
Adams, R.B., Licht, A.N., Sagiv, L., 2011. Shareholders and stakeholders: How do directors decide? Strateg. Manag. J. 32 (12), 1331–1355.
Afzali, M., Khan, U., Rajgopal, S., 2022. Sharing the pain between workers and management: Evidence from the COVID-19 pandemic and 9/11 attacks. Available
at SSRN: https://ssrn.com/abstract=4053005.
Alekseev, G., Amer, S., Gopal, M., Kuchler, T., Schneider, J.W., Stroebel, J., Wernerfelt, N., 2023. The effects of COVID-19 on U.S. small businesses: Evidence
from owners, managers, and employees. Manage. Sci. 69 (1), 7–24.
Bartik, A., Bertrand, M., Lin, F., Rothstein, J., Unrath, M., 2020. Measuring the labor market at the onset of the COVID-19 crisis. Working Paper 27613, National
Bureau of Economic Research.
Bayat, A., Goergen, M., 2022. CEO political ideology, shareholder primacy and dividend policy. European Corporate Governance Institute – Finance Working
Paper. Available at SSRN: https://ssrn.com/abstract=3522270.
Benartzi, S., Michaely, R., Thaler, R., 1997. Do changes in dividends signal the future or the past? J. Finance 52 (3), 1007–1034.
Benton, R.A., Cobb, J.A., Werner, T., 2022. Firm partisan positioning, polarization, and risk communication: Examining voluntary disclosures on COVID-19.
Strateg. Manag. J. 43 (4), 697–723.
Bizjak, J.M., Kalpathy, S.L., Mihov, V.T., Ren, J., 2022. CEO political leanings and store-level economic activity during COVID-19 crisis: Effects on shareholder
value and public health. J. Finance 77, 2949–2986.
BLS, 2012. The Recession of 2007–2009. Bureau of Labor Statistics, Washington DC.
Campbell, K., Goenner, C.F., Notbohm, M., Smedema, A., 2022. Political ideology and CEO performance under crisis. Rev. Quant. Financ. Account. 58 (1),
329–359.
Campello, M., Graham, J.R., Harvey, C.R., 2010. The real effects of financial constraints: Evidence from a financial crisis. J. Financ. Econ. 97 (3), 470–487.
Canes-Wrone, B., Rothwell, J.T., Makridis, C., 2020. Partisanship and policy on an emerging issue: Mass and elite responses to COVID-19 as the pandemic
evolved.
Carhart, M.M., 1997. On persistence in mutual fund performance. J. Finance 52 (1), 57–82.
Cejnek, G., Randl, O., Zechner, J., 2021. The COVID-19 pandemic and corporate dividend policy. J. Financ. Quant. Anal. 56 (7), 2389–2410.
Chetty, R., Friedman, J., Hendren, N., Stepner, M., Team, T.O.I., 2022. The economic impacts of COVID-19: Evidence from a new public database built using
private sector data. Working Paper 27431, National Bureau of Economic Research.
Chin, M.K., Hambrick, D.C., Treviño, L.K., 2013. Political ideologies of CEOs: The influence of executives’ values on corporate social responsibility. Adm. Sci. Q.
58 (2), 197–232.
Chin, M.K., Zhang, S.X., Jahanshahi, A.A., Nadkarni, S., 2021. Unpacking political ideology: CEO social and economic ideologies, strategic decision-making
processes, and corporate entrepreneurship. Acad. Manag. J. 64 (4), 1213–1235.
Congdon, W.J., Vroman, W., 2022. The Unemployment Insurance System in Two Recent Economic Downturns: Lessons from the Great Recession and the COVID-19
Recession. Report Prepared for the Us Department of Labor (dol), Urban Institute, Washington, DC.
Cookson, J.A., Engelberg, J.E., Mullins, W., 2020. Does partisanship shape investor beliefs? Evidence from the COVID-19 pandemic. Rev. Asset Pricing Stud. 10
(4), 863–893.
De Angelis, D., Grinstein, Y., 2020. Relative performance evaluation in CEO compensation: A talent-retention explanation. J. Financ. Quant. Anal. 55 (7),
2099–2123.
DeAngelo, H., DeAngelo, L., Skinner, D.J., 1992. Dividends and losses. J. Finance 47 (5), 1837–1863.
DeAngelo, H., DeAngelo, L., Skinner, D.J., 2000. Special dividends and the evolution of dividend signaling. J. Financ. Econ. 57 (3), 309–354.
Dechow, P.M., Erhard, R.D., Sloan, R.G., Soliman, M.T., 2021. Implied equity duration: A measure of pandemic shutdown risk. J. Account. Res. 59 (1), 243–281.
Di Giuli, A., Kostovetsky, L., 2014. Are red or blue companies more likely to go green? Politics and corporate social responsibility. J. Financ. Econ. 111 (1),
158–180.
Donangelo, A., Gourio, F., Kehrig, M., Palacios, M., 2019. The cross-section of labor leverage and equity returns. J. Financ. Econ. 132, 497–518.
Elnahas, A.M., Kim, D., 2017. CEO political ideology and mergers and acquisitions decisions. J. Corp. Finance 45, 162–175.
Fama, E.F., French, K.R., 1993. Common risk factors in the returns on stocks and bonds. J. Financ. Econ. 33 (1), 3–56.
Federal Reserve, 2020. Assessment of Bank Capital during the Recent Coronavirus Event. Board of Governors of the Federal Reserve System, Washington, DC.
Francis, B.B., Hasan, I., Sun, X., Wu, Q., 2016. CEO political preference and corporate tax sheltering. J. Corp. Finance 38, 37–53.
Graham, J.R., Kim, H., Leary, M., 2020. CEO-board dynamics. J. Financ. Econ. 137 (3), 612–636.
Guenzel, M., Hamilton, C., Malmendier, U., 2023. CEO social preferences and layoffs. Working paper.
Gupta, A., Briscoe, F., Hambrick, D.C., 2018. Evenhandedness in resource allocation: Its relationship with CEO ideology, organizational discretion, and firm
performance. Acad. Manag. J. 61 (5), 1848–1868.
Healy, P.M., Palepu, K.G., 1988. Earnings information conveyed by dividend initiations and omissions. J. Financ. Econ. 21 (2), 149–175.
Hutton, I., Jiang, D., Kumar, A., 2014. Corporate policies of Republican managers. J. Financ. Quant. Anal. 49 (5–6), 1279–1310.
IMF, 2020. World economic outlook, April 2020. International Monetary Fund.
Iyer, S.R., Feng, H., Rao, R.P., 2016. Payout flexibility and capital expenditure. Rev. Quant. Financ. Account. 49 (3), 633–659.
Jensen, G.R., Lundstrum, L.L., Miller, R.E., 2010. What do dividend reductions signal? J. Corp. Finance 16 (5), 736–747.
Kaplan, S.N., Reishus, D., 1990. Outside directorships and corporate performance. J. Financ. Econ. 27 (2), 389–410.
Keum, D.D., Meier, S., 2023. License to layoff? Unemployment insurance and the moral cost of layoffs. Organ. Sci. 35 (3), 994–1014.
Levenshtein, V.I., 1966. Binary codes capable of correcting deletions, insertions and reversals. Sov. Phys. Doklady 10 (8), 707–710.
Lintner, J., 1956. Distribution of incomes of corporations among dividends, retained earnings, and taxes. Amer. Econ. Rev. 46 (2), 97–113.
Matsa, D.A., Miller, A.R., 2014. Workforce reductions at women-owned businesses in the United States. ILR Rev. 67 (2), 422–452.
Michaely, R., Thaler, R.H., Womack, K.L., 1995. Price reactions to dividend initiations and omissions: Overreaction or drift? J. Finance 50 (2), 573–608.
Nickell, S., 1984. An investigation of the determinants of manufacturing employment in the United Kingdom. Rev. Econ. Stud. 51 (4), 529–557.
Parrino, R., Sias, R.W., Starks, L.T., 2003. Voting with their feet: Institutional ownership changes around forced CEO turnover. J. Financ. Econ. 68 (1), 3–46.
Pennington, J., Socher, R., Manning, C.D., 2014. Glove: Global vectors for word representation. In: Proceedings of the 2014 Conference on Empirical Methods
in Natural Language Processing. EMNLP, pp. 1532–1543.
Pettifor, A., 2006. The coming first world debt crisis. Palgrave Macmillan, UK.
Rothwell, J., Makridis, C., 2020. Politics is Wrecking America’s Pandemic Response. Brookings Institution, Washington DC, https://www.brookings.edu/articles/
politics-is-wrecking-americas-pandemic-response/.
Schaeck, K., Cihak, M., Maechler, A., Stolz, S., 2011. Who disciplines bank managers? Rev. Finance 16 (1), 197–243.
Unsal, O., Hassan, M.K., Zirek, D., 2016. Corporate lobbying, CEO political ideology and firm performance. J. Corp. Finance 38, 126–149.
Vorsatz, B., 2022. Costs of political polarization: Evidence from mutual fund managers during COVID-19. European Corporate Governance Institute – Finance
Working Paper. Available at SSRN: https://ssrn.com/abstract=3734026.
WHO, 2021. WHO-convened global study of origins of SARS-CoV-2: China part. Available at: https://www.who.int/publications/i/item/who-convened-global-
study-of-origins-of-sars-cov-2-china-part.
Yonker, S.E., 2017. Do managers give hometown labor an edge? Rev. Financ. Stud. 30 (10), 3581–3604.
28
File and source
- File
- the-impact-of-ceo-political-ideology-on-labor-cost-reductions-and-payout-decisions-during-the-covid-19-pandemic.pdf
- Size
- 1,774,031 bytes
- SHA-256
- 02db3dcad58ad748e83bec0d55bd4c1feed001857d68772472300ccc9e9fffa1
- Original
- doi.org