How COVID-19 affected corporate dividend decisions: Novel evidence from emerging countries

AlGhazali, Abdullah; Yilmaz, Ilker Article How COVID-19 affected corporate dividend decisions: Novel evidence from emerging countries Financial Internet Quarterly Provided in Cooperation with: University of Information Technology and Management, Rzeszów Suggested Citation: AlGhazali, Abdullah; Yilmaz, Ilker (2023) : How COVID-19 affected corporate dividend decisions: Novel evidence from emerging countries, Financial Internet Quarterly, ISSN 2719-3454, Sciendo, Warsaw, Vol. 19, Iss. 4, pp. 25-48, https://doi.org/10.2478/fiqf-2023-0025 This Version is available at: https://hdl.handle.net/10419/329856 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/ 10.2478/fiqf-2023-0025 Abstract The study aims to investigate the corporate dividend policy decisions in emerging countries during the COVID-19 pandemic. Our sample consists of 5,869 publicly listed firms from 29 emerging countries to explicate the observed trends in dividend policy during the pandemic. Logistic regressions are used to investigate the main factors that drive the propensity to change dividend payouts. Our analysis reveals that most firms opted to either increase or decrease their dividends, with a minority proportion deciding to maintain dividends. Notably, our findings demonstrate that firm profitability is the main driver of all types of dividend changes, except when firms opt to maintain or decrease dividends. Moreover, we find that when firms reduce dividends by over 70%, profitability emerges as a crucial determinant, thus bolstering the signaling hypothesis. The results are robust to sample size sensitivity and different levels of dividend changes. The findings of the study might have practical implications for corporate managers and policymakers in designing dividend decisions and policies under uncertain conditions. This research underscores the impact of the COVID-19 pandemic on corporate dividend policy in emerging countries and emphasizes the need to consider the level of dividend changes in exploring the dividend puzzle. JEL classification: G30, G32, G35 Keywords: Dividend policy, COVID-19, Pandemic, Profitability, Emerging countries Received: 01.06.2023 Accepted: 08.09.2023 Cite this: AlGhazali M.A. & Yilmaz I. (2023) How COVID-19 Affected Corporate Dividend Decisions: Novel Evidence from Emerging Countries. Financial Internet Quarterly 19(4), pp. 25-48. © 2023 Abdullah Mohammed AlGhazali and Ilker Yilmaz, published by Sciendo. This work is licensed under the Creative Commons Attribution- NonCommercial-NoDerivatives 3.0 License. 1 Dhofar University, Department of Finance and Economics, Oman, e-mail: aalghaz[email protected]du.om, ORCID: https://orcid.org/0000-0002-6378-3044. 2 Dhofar University, Department of Accounting, Oman, e-mail: iy[email protected]du.om, ORCID: https://orcid.org/0000-0001-8501-9664. The agency theory of dividends, proposed by Jensen (1986), contends that firms with excess cash can resolve the principal-agent problem by maintaining or increasing dividends. In contrast, the signaling theory of dividends, advanced by Bhattacharya (1979), asserts that changing dividends can convey valuable information about a firm's prospects. In this context, our objective is to investigate the corporate dividend policies during the COVID-19 pandemic, utilizing a substantial sample from emerging countries. Our study seeks to provide new insights into the effect of the pandemic on corporate dividend policies in emerging countries, where empirical research in this domain is relatively limited. To the best of our knowledge, our study is the first of its kind to explore the impact of the pandemic on corporate dividend policy across 29 emerging countries. The present study represents a significant contribution to the literature on corporate dividend policy. Specifically, it is the first study to investigate the impact of unexpected exogenous shocks, such as the COVID-19 pandemic, on dividend policy using a large sample from emerging countries. The empirical findings of our study demonstrate that the majority of firms in these countries either increase or cut dividends during the pandemic. Additionally, the number of firms that maintain dividends is lower than those that omit dividends, which underscores the impact of the pandemic on firms' dividend stability in these countries. Our study also sheds light on the importance of considering the levels of dividend changes to explain the variation in dividend policy across firms and countries. The results reveal that at a higher level of dividend reduction (> 75%), there is a significant negative growth in the profitability of firms that decrease dividends compared to those that maintain dividends. Furthermore, at a higher level of dividend increases and decreases (> 75%), profitability and asset turnover are the primary drivers of corporate dividend decisions. However, the decision to increase or maintain a dividend is primarily attributed to the profitability and size of a firm. Therefore, splitting dividend changes into levels may provide further insights into the mixed evidence on corporate dividend policy. Lastly, our study highlights the variation in dividend policy between developing (Ali, 2022) and developed countries, which merits further consideration. These findings have significant implications for policymakers, investors, and other stakeholders, particularly regarding the impact of unexpected exogenous shocks on dividend policy in developing economies. Overall, our study makes a valuable contribution to the literature on corporate finance and dividend policy, and its findings have important implications for future research in this area. COVID-19 has been one of the most challenging uncertainties for corporations in recent years. It was first detected in December 2019, and its effects became more concrete by February 2020. The World Health Organization (WHO) officially declared it a pandemic in March 2020. To cope with this pandemic, governments announced several actions to prevent the virus's spread, including partial or complete lockdowns, which led to significant effects on economies and corporations. This resulted in declines in economic growth worldwide by 4.2%, in the US by 3.4%, in Europe by 7.5%, in the G-20 by 3.8%, and in India by 9.9%. Yilmazkuday (2020) reports the negative impact of COVID -19 on the global economy, while Mazur et al. (2021), Tripathi and Pandey (2021), and Baker et al. (2020) demonstrate a robust increase in equity market volatility in the US. The uncertainty caused by COVID-19 has affected various industries and regions in disparate ways. Pan (2021) has reported a significant drop in manufacturing PMI since 2011, with the effect being more pronounced in developed rather than developing markets. However, the MSCI emerging markets index has underperformed the MSCI world index during the pandemic (Pan, 2021). This effect has extended to internal and external capital chains, prompting firms to review their financial policies (Jiang et al., 2021). During 2020, leverage decreased significantly in the US (Haque & Varghese, 2021). Firms with high leverage experience a high level of risk (Huang & Ye, 2021). The flow of credit to industrial sectors remained robust (Deghi et al., 2021), and stock markets have reacted negatively to the pandemic (Harjoto et al., 2021; Prabheesh et al., 2020). Wigglesworth et al. (2020) reported a reduction in firms' dividends globally during the pandemic. The emerging empirical research on the impact of COVID- 19 on corporate dividend policy has focused on developed countries (e.g. Ali, 2022; Ntantamis & Zhou, 2022). However, few studies have explored its effects on emerging countries and have used a single-country setting, such as Ali et al. (2022) in Pakistan, and Tinungki et al. (2022) in Indonesia. Scholars have demonstrated salient differences between developed and developing countries regarding corporate governance mechanisms, legal protection, voting rights, ownership structure, and the role of institutional shareholders (Glen et al., 1995; Mitton, 2004; Adjaoud & Ben-Amar, 2010). These issues are not independent of how corporate dividend policy is determined and deserve further investigation. For instance, Aivazian et al. (2003) have demonstrated that the sensitivity of dividend policy determinants differs in emerging compared to developed markets. of this crisis on dividend policy (e.g. Hauser, 2013; Floyd et al., 2015). For instance, Hoberg and Prabhala (2009) detect a lower propensity of firms to pay dividends after the financial crisis, and Hauser (2013) finds consistent results with this conjecture. Bistrova et al. (2013) show that there was a reduction in the payout policy during the financial crisis in European firms. COVID-19 has been a similar turmoil period for corporations, and they have encountered financial policy challenges, including dividend decisions (Cejnek et al., 2021; Ali, 2022; Eugster et al., 2022; Ntantamis & Zhou, 2022). Ali (2022) investigates the impact of COVID-19 on dividend policy in G-12 countries and finds that while the majority of firms maintain or increase dividends, there is a significant increase in the number of firms that decrease or omit dividends compared to the pre- COVID-19 period. Her findings reveal that firms' profitability plays a crucial role in determining the decision to change dividends. Using US data, Krieger et al. (2021) study the impact of COVID-19 on the payout policy of US firms, reporting that the proportion of dividend cuts or omissions during 2020 was three to five times higher than in the periods 2015-2019. Ntantamis and Zhou (2022) examine the effect of COVID-19 on the payout policies of firms in G-7 countries, considering dividends and share repurchases. They find that more companies decreased their payout after the pandemic started and point out that the scale of adjustments varies across countries. They also find that cash holdings helped mitigate the negative effects of the pandemic, with the effect being more significant in North America and Japan compared to Europe. In developing countries, Tinungki et al. (2022) examine the impact of COVID-19 on dividend policy in Indonesia and find that the pandemic does not have a significant effect on firms' dividend policy. However, Ali et al. (2022) demonstrate that the majority of listed firms in Pakistan omit dividends during the pandemic, while firms that decide to maintain dividends account for less than 6% of the sample. They further show that firms that increase (decrease) dividends experience a positive (negative) profitability compared with firms that decrease (maintain) dividends. However, they find no robust evidence on other dividend change groups. The study's sample comprises listed firms from 29 countries that were obtained from Refinitiv Eikon during the 2015-2020 period. We follow the recent studies that examine the impact of COVID-19 on dividends and choose the period 2015-2020 (e.g. Krieger et al., 2020; Ali, 2022). The initial sample consisted of This paper is structured as follows: Section 2 provides a review of the relevant literature, while Section 3 describes the data and methodology used in our analysis. In Section 4, we present our empirical results, and Section 5 reports on the robustness checks we conducted. Finally, Section 6 concludes the paper. Dividend policy is one of the financial policy challenges faced by corporations. Some studies have focused on the question of whether dividend policy affects firm value, while others have focused on the determinants of dividend policy. In their seminal work, Miller and Modigliani (1961) argue that in a perfect market, corporate dividend policy is irrelevant and does not have any impact on corporate value. However, in the real world with market imperfections such as taxes, transaction costs, asymmetric information, and principal-agent conflict, dividend policy has been shown to affect shareholders' value. The existing literature documented that there are significant differences in dividend policies and decisions of the firms in emerging countries and developed countries, particularly the firms in emerging countries follow less stable dividend policies and the most important determinant of the dividend decision is the current year earnings, also the firms in emerging countries are subject to higher financial constraints (Adaoglu, 2000; Aivazian et al., 2003; Glen & Singh, 2004). Jabbouri (2016) investigated the determinants of dividend policy in MENA region countries and reported that firm size, profitability, and liquidity have a positive effect on dividend payments while firm growth and leverage have a negative effect. The responses of the firms in their dividend policies during economic slumps are also different in emerging and developed countries. The firms in developed countries tend to reduce dividends in such periods while the counterparts in emerging countries tend to increase the payout to pacify the investors (Chemmanur & Tian, 2014; Jabbouri, 2016). Agency and signaling theories have been widely used in the literature to justify the relevance of corporate dividend policy. Agency theory explains dividend decisions in principal-agent problems (Jensen, 1986). In this context, firms should continue to pay or increase dividends to prevent self-interested managers from investing excess cash in negative NPV projects or obtaining private benefits. The signaling theory argues that dividend changes convey signals about firms' prospects, suggesting a positive link between dividends and earnings (Bhattacharya, 1979). Several empirical studies have examined corporate dividend policy during the financial crisis of 2007-2009 and provided empirical evidence of the adverse impact of potential outliers, we implemented a winsorization procedure on all non-dummy variables, limiting extreme values to the 1st and 99th percentiles. This technique effectively mitigates the impact of any errant observations, thereby promoting a more robust and reliable dataset for subsequent analyses. As a result, our final sample comprised 5,869 firms from 29 countries. We utilized the Industry Classification Benchmark (ICB) to categorize firms into nine distinct groups, as reported in Table 1.3 14,208 firms, from which 738 financial and real estate firms were removed. We excluded firms that never paid dividends or engaged in share repurchases in 2020 (N = 6,738) from the sample. Additionally, we removed firms that chose to omit dividends in 2019 or initiate dividends only in 2020, following Ali's (2022). We retained only firms with complete accounting data and restricted the sample to investable firms by excluding those with total assets and total equity of less than 0.5 and 0.25 million, respectively. To counter the influence Table 1: Sample Details Panel A: Sample distribution per country Country Freq. Percent Cum. Country Freq. Percent Cum. Argentina 22.0 0.4 0.4 Morocco 22.0 0.4 80.7 Bahrain 13.0 0.2 0.6 Oman 39.0 0.7 81.4 Bangladesh 69.0 1.2 1.8 Pakistan 145.0 2.5 83.9 Brazil 120.0 2.0 3.8 Peru 55.0 0.9 84.8 Bulgaria 14.0 0.2 4.1 Philippines 60.0 1.0 85.8 Chile 88.0 1.5 5.6 Poland 90.0 1.5 87.3 China 2620.0 44.6 50.2 Qatar 19.0 0.3 87.7 Colombia 29.0 0.5 50.7 Romania 41.0 0.7 88.4 Egypt 54.0 0.9 51.6 Russia 63.0 1.1 89.4 Hungary 6.0 0.1 51.7 Saudi Arabia 50.0 0.9 90.3 India 1085.0 18.5 70.2 South Africa 90.0 1.5 91.8 Indonesia 172.0 2.9 73.1 Thailand 389.0 6.6 98.5 Kuwait 20.0 0.3 73.5 Turkey 72.0 1.2 99.7 Malaysia 358.0 6.1 79.6 UAE 19.0 0.3 100.0 Mexico 45.0 0.8 80.3 Total 5869.0 100.0 Panel B: Sample distribution per industry ICB Industry name Freq. Percent Cum. ICB Industry name Freq. Percent Cum. Basic Materials 929.0 15.8 15.8 Industrials 1553.0 26.5 82.7 Consumer Discretionary 1089.0 18.6 34.4 Technology 524.0 8.9 91.7 Consumer Staples 630.0 10.7 45.1 Telecommunications 166.0 2.8 94.5 Energy 187.0 3.2 48.3 Utilities 323.0 5.5 100.0 Health Care 468.0 8.0 56.3 Total 5869.0 100.0 Source: Author’s own work. hibiting a decrease (or no-change) in dividends represent the second (or third) largest group. Additionally, in 2020, the number of firms with dividend increase (DIC) stood at 2,353, surpassing all other types of dividendschange groups, a trend that aligns with Ali's (2022) observations in G-12 countries. However, this pattern has remained relatively flat since 2019, diverging from that observed in developed countries. Table 2 provides a comprehensive overview of summary statistics by dividend-change groups for the 2015-2020 period. Notably, the vast majority of firms in markets have exhibited a propensity for a dividend increase, aligning with the findings reported in extant research conducted on developed markets (Ali, 2022). However, contrary to these previous studies, firms ex- ……………………………………………………………………………………………………………………………………………………………………………. 3 The majority of firms in Table 1 are from China and India which account for 63% of the sample. This might lead our estimations to be biased. Hence, we consider the overrepresentation of the sample in the robustness section. dividends (Panels D). All variables used in the analysis are defined in Appendix A. Among the different dividend-change groups, the firms that increased dividends were found to be more profitable and larger, this is consistent with the findings reported in developed countries (Ali, 2022). The firms that decided not to change dividends were observed to be more liquid during the COVID-19 year. However, their profitability, assets turnover, size, and market-to-book ratio were found to be very similar to the dividend-decreasing firms, which contradicts the findings of Ali (2022) in G-12 countries. These results suggest that there are similarities in the characteristics of firms that maintain dividends and those that cut dividends, which is not in line with Ali's (2022) findings that show that firms that decide not to change dividends are much more profitable, have higher assets turnover, are smaller, and experience lower market-to-book ratios. On the other hand, the dividend-omitting firms were found to have negative profitability, lower liquidity, more debt, and a high market-to-book ratio. Table 2 shows that the number of firms that opted not to change dividends (DNC) in 2020 stood at 871, in contrast to the 1,125 recorded the previous year. This finding contradicts that reported in G-12 nations (Ali, 2022), where the number of firms that maintained dividends in 2019 and 2020 was almost identical. Dividenddecreasing firms (DDC) remained relatively stable both during the pandemic and preceding years, diverging from the sample observed in developed countries (Ali, 2022). Conversely, dividend-omitting firms (DOM) increased over the period, reaching their highest levels during the pandemic year, a trend that aligns with the results of research conducted in the US (Pettenuzzo et al., 2021) and developed countries (Ali, 2022). For the remainder of this study, we will focus on the pandemic year: 2020.4 Table 3 presents the descriptive statistics for each dividend group during the pandemic year, which includes firms that increased dividends (Panels A), firms that maintained dividends (Panels B), firms that decreased dividends (Panels C), and firms that omitted Table 2: Number of Firms per Dividend-Change Group 2015 2016 2017 Total Dividend Obs % Obs % Obs % Obs DIC 1571 42.3 2294 45.5 2605 47.3 13775 DNC 557 15.0 946 18.8 1026 18.6 5545 DDC 1104 29.7 1277 25.3 1364 24.8 8584 2018 2019 2020 Total Dividend Obs % Obs % Obs % Obs DIC 2670 44.4 2282 38.6 2353 40.1 13775 DNC 1020 17.0 1125 19.0 871 14.8 5545 DDC 1562 26.0 1652 27.9 1625 27.7 8584 DOM 763 12.7 860 14.5 1020 17.4 4165 Total 6015 100.0 5919 100.0 5869 100.0 32069 DOM 483 13.0 525 10.4 514 9.3 4165 Total 3715 100.0 5042 100.0 5509 100.0 32069 DIC: Dividend increase, DDC: Dividend decrease, DNC: Dividend no change, DOM: Dividend omissions. Source: Author’s own work. …………………………………………………………………………………………………………………………………………………………………………… 4 The total number of firms paying dividends in our study, including DIC, DDC, and DNC, decreased from 5,059 to 4,849 in 2019 and 2020, respectively. This finding is consistent with the results obtained by Ntantamis and Zhou (2022) in G-7 countries. However, the numbers in each dividend-change group differ significantly from those reported in developed countries (Ali, 2022), indicating differences in dividend behavior between developing and developed markets. tion that they have a positive correlation with liquidity. Dividend cuts have a positive association with ROA, ROE, operating profit margin, and size, but are negatively associated with changes in earnings, asset turnover, leverage, liquidity, and market-to-book ratio. Dividend omissions display an opposite pattern compared to the other three categories; they have negative correlations with all profitability measures, asset turnover, Table 4 (see: Appendix) presents the pairwise correlations among the variables used in the analyses during the COVID-19 period. The dividend increases are positively correlated with all profitability measures, asset turnover, and size, but negatively correlated with leverage, liquidity, and market-to-book ratio. Dividend no-change cases have a similar pattern, with the excep- Table 3: Descriptive statistics Characteristics Mean Median Min Max Std. Dev. Panel A: Dividend increases (DIC) ROA % 7.8 6.6 -34.0 27.5 5.8 chE % 3.2 2.5 -87.2 82.2 9.1 ROE % 14.2 12.2 -145.7 61.8 11.2 Operpm % 15.8 12.8 -267.1 65.6 14.4 AstTvr 0.8 0.7 0.0 3.5 0.5 Lev % 43.1 42.6 3.0 94.0 19.5 Size 20.1 20.1 14.6 23.9 1.8 Operpm % 10.3 9.3 -267.1 65.6 16.2 AstTvr 0.8 0.7 0.0 3.5 0.6 Lev % 41.6 41.0 3.0 94.0 19.9 Size 19.6 19.6 14.4 23.9 1.8 Liq 2.9 1.8 0.2 25.5 3.4 MktBk 0.8 0.6 0.0 9.6 0.8 Panel C: Dividend decreases (DDC) ROA % 4.8 3.9 -34.0 27.5 5.3 chE % -4.3 -2.5 -109.8 90.1 11.9 ROE % 9.1 7.2 -85.6 61.8 11.3 Operpm % 10.9 9.0 -192.5 65.6 16.0 AstTvr 0.7 0.6 0.0 3.5 0.5 Lev % 43.5 43.8 3.0 94.0 20.8 Size 19.9 19.8 14.7 23.9 1.8 Liq 2.6 1.7 0.2 25.5 2.8 MktBk 0.8 0.6 0.0 9.6 1.0 Panel D: Dividend omissions (DOM) ROA % -0.4 0.4 -34.0 27.5 8.2 chE % -11.3 -6.2 -109.8 90.1 21.0 ROE % -2.9 0.9 -145.7 61.8 22.7 Operpm % -3.5 2.6 -267.1 65.6 33.7 AstTvr 0.7 0.6 0.0 3.5 0.6 Lev % 47.8 48.3 3.0 94.0 20.9 Size 19.2 19.3 14.4 23.9 1.9 Liq 2.3 1.5 0.2 25.5 3.1 MktBk 1.1 0.7 0.0 9.6 1.3 The table presents several characteristics of the sample. It reports the mean, median, maximum, minimum and standard deviation of variables for each dividend’s category. All variables are defined in Appendix A. Panels A, B, C, and D present the groups of firms that chose to increase, not change, decrease and omit dividends, respectively. Source: Author’s own work. Panel B: No change in dividends (DNC) ROA% 5.3 4.5 -32.1 27.5 5.0 chE% -1.2 -0.1 -109.8 50.4 8.4 ROE% 9.2 8.6 -145.7 46.7 10.2 Liq 2.6 1.8 0.2 25.5 2.6 MktBk 0.7 0.4 0.0 9.6 0.9 ov, 2008; Ali, 2022). All other variables are defined in Appendix A. Furthermore, we control for country and industry fixed effects in all regressions. The estimates for the logistic regression are displayed in Table 5 (see: Appendix), Panels A to D. Considering the results in Panels A to E (Models 1 to 16), we found strong associations between firms' profitability and the propensity to change dividends in emerging countries during the COVID-19 pandemic, which is in line with Ali's recent study (2022) in G-12 countries. Panel A reveals that profitability measures are positive and statistically significant at the 1% level, indicating that firms with higher profitability are more likely to increase dividends than to maintain them. Models 5 to 8 in Panel B document that the likelihood of increasing dividends is more pronounced in firms with higher profitability, as opposed to decreasing them. Firms with lower profitability, as stated in models 9 to 12 of Panel C, are more likely to omit dividends than to maintain their levels. Moreover, Panel D detects that lower profitability increases the likelihood of firms omitting dividends rather than decreasing them. The regression outputs in models 17 to 20 demonstrate that the coefficients of profitability measures are not robustly significant. These findings provide little support for the impact of profitability on the likelihood of firms decreasing dividends compared to maintaining them, which is inconsistent with the findings in developed countries (Ali, 2022). Study (2022) in G-12 countries. Panel A reveals that profitability measures are positive and statistically significant at the 1% level, indicating that firms with higher profitability are more likely to increase dividends than to maintain them. Models 5 to 8 in Panel B document that the likelihood of increasing dividends is more pronounced in firms with higher profitability, as opposed to decreasing them. Firms with lower profitability, as stated in models 9 to 12 of Panel C, are more likely to omit dividends than to maintain their levels. Moreover, Panel D detects that lower profitability increases the likelihood of firms omitting dividends rather than decreasing them. The regression outputs in models 17 to 20 demonstrate that the coefficients of profitability measures are not robustly significant. These findings provide little support for the impact of profitability on the likelihood of firms decreasing dividends compared to maintaining them, which is inconsistent with the findings in developed countries (Ali, 2022). and liquidity, but positive correlations with leverage, size, and market-to-book ratio. These findings reveal that the no-change dividend and dividend decrease groups exhibit similar correlations with all of the used variables, except for ChE, size, and liquidity. These statistics confirm some differences from those in developed markets (Ali, 2022), which show that firms that decrease dividends are negatively correlated with all profitability measures, firm size, and market-to-book ratio. To examine the impact of the pandemic on corporate dividend policy, we follow the recent study by Ali (2022). We calculate the dividend changes following Nissim (2001), as the difference between dividends in fiscal year t and the dividends in the previous year, scaled by the dividend in the previous year. Then, dividend changes are sorted into four groups: (I) dividend increases, (II) dividend no changes, (III) dividend decreases, and (IV) dividend omission. Next, dichotomous variables are constructed based on each two groups of dividend changes (DivChange). Our dependent variable is a categorical variable that equals 1 or 0. Thus, we apply logistic regression to investigate what factors drive the variation in dividend decisions in emerging countries during the COVID-19 pandemic. The model is specified as follows: (1) Where is a dichotomous variable that takes the value of 1 and 0 for each two groups: dividend increases (= 1) versus dividend no change (= 0); dividend increases (= 1) versus dividend decreases (= 0); dividend omissions (= 1) versus dividend no changes (= 0); dividend omission (= 1) versus dividend decreases (= 0); and dividend decreases (= 1) versus dividend no change (= 0). We have used four different measures of profitability following the recent literature (i.e. Return on assets (ROA %); Change in earnings (chE %); Return on equity (ROE %); Operating profit margin (Operpm %)). ROA % is defined as net income over total assets (Krieger et al., 2021), and chE % is defined as the change in the net income scaled by the book value of equity (Ali, 2020). ROE % is defined as net income scaled by the book value of equity (Richard et al., 2014). Operpm % is operating profit divided by revenue (Fairfield & Yohn, 2001).5 Control variables, include assets turnover, firm size, leverage, liquidity, and market-to-book ratio (DeAngelo et al., 2004; Denis & Osob- 01 ( 1) i i i j k i Pr DivChange Profitabitliy X Industry Dummies Country Dummies     = = + +  + ++ ………………………………………………………………………………………………………………………………………………………………………….. 5 We have employed different measures of profitability to provide robust evidence of the impact of profitability on corporate dividend policy. The majority of the previous studies have demonstrated the significant influence of profitability on corporate dividend policy (e.g. Fama & French, 2001; DeAngelo et al., 2004; Al-Ghazali, 2014). the propensity of firms to change dividends and profitability diminishes at moderate and high levels of dividend reductions (25% ≤ DDCD < 50% and 50% ≤ DDCD < 75%) as reported in Panel B and C. Specifically, we find that at these levels of dividend reductions, the profitability measures are not robustly significant. Assets turnover is negative and statistically significant in panel B indicating that firms at moderate levels of dividend reduction exhibit lower assets turnover than those that maintain dividends. Panel C shows that size and liquidity bear statistically negative coefficients. In the case of extreme dividend reduction (a decrease ≥ 75%), as in Panel D, the findings demonstrate a robust significant negative correlation between all the profitability measures and the propensity of firms to decrease rather than maintain dividends, consistent with developed markets (Ali, 2022). Furthermore, firms with extreme dividend reductions exhibit lower asset turnover than those that maintain dividends. In the preceding section, we presented compelling evidence of the impact of the COVID-19 pandemic on the dividend policies of corporations in nations except in one group: Dividend decreases vs. dividend nochange. Nonetheless, it is plausible that our findings are attributable to alternative explanations. To fortify our results, we address two critical factors in this section: (1) the sensitivity of sample size; and (2) the distinction between the levels of dividend increases and the maintenance of existing dividend levels. The present study encompasses data from 29 distinct countries, albeit with variations in the number of observations for each country, as indicated in Table 1. The preponderance of data from China and India in our sample warrants scrutiny, as this may introduce a potential bias into our estimation through overrepresentation. To address this concern, we re-examine our analysis, as reported in Table 5 (see: Appendix), by omitting data from the aforementioned countries. The estimated outputs from this refined analysis are subsequently presented in Table 7 (see: Appendix). Our findings, which align with those reported in Table 5 (see: Appendix), furnish compelling evidence of the impact of profitability measures on corporate dividend policy, except in one category, i.e., dividend decreases vs. dividend no-change. Moreover, the remaining estimated coefficients demonstrate consistent signs and levels of significance. In light of these findings, we affirm that our estimations remain robust, notwithstanding the potential for overrepresentation in our sample. The effect of firms’ characteristics on the likelihood of firms to change dividends in panel A of Table 5 (see: Appendix) shows that the propensity of firms to increase rather than maintain dividends is positively associated with assets turnover as shown in models 2 and 3, indicating that firms with high assets turnover are more likely to increase dividends. Size bears positive and significant coefficients indicating that larger firms are more likely to increase than maintain dividends. Furthermore, the coefficients of liquidity are statistically insignificant at 10%, suggesting that firms that increase compared to those that maintain dividends do not exhibit significant liquidity differences. The market- to-book ratio is lower in firms that increase rather than maintain dividends. As shown in models 5 to 8 of panel B, the likelihood of firms to increase than decrease dividends is positively (negatively) correlated with assets turnover, size, and market-to-book ratio. Panels C and D reveal that the propensity of firms to omit rather than maintain dividends (penal C) and omit rather than decrease dividends (panel D) is negatively (positively) and significantly related to assets turnover and size (leverage and market-to-book ratio). Panel E reports that asset turnover reduces the propensity of firms to decrease rather than maintain dividends while other factors are not statistically significant. We extended our analysis to investigate the inconsistent results with Ali's study on the impact of firms' profitability on the likelihood of firms decreasing rather than maintaining dividends. We divided dividend reductions into four groups: (I) reduction less than 25%; (II) reduction between 25% and less than 50%; (III) reduction between 50% and less than 75%; and (IV) reduction greater than 75% and less than 100%. We ran a logistic regression using Eq. (1), where the explanatory variables are (I) a dichotomous variable that is 1 for dividend reduction less than 25% and 0 if dividends are not changed; (II) a dichotomous variable that is 1 for dividend reduction between 25% and less than 50% and 0 if dividends are not changed; (III) a dichotomous variable that is 1 for dividend reduction between 50% and less than 75% and 0 if dividends are not changed; and (IV) a dichotomous variable that is 1 if dividend reduction between 75% and less than 100% and 0 if dividends are not changed. The estimated outputs of the logistic regression are presented in Table 6 Panel A to D (see: Appendix). We find strong evidence indicating that at a lower level of dividend reduction, Panel A, higher profitable firms are more likely to cut than maintain dividends. The coefficients of leverage and size are positive and significant only in models 1 and 2, respectively. These findings might suggest that a small reduction in dividends could be used by firms not to convey their prospect about future profitability: signaling. The relationship between Table 5 (A): Dividend Changes during COVID-19 Panel A: DICD vs. DNCD Panel B: DICD vs. DDCD Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 ROA 0.1310*** 0.1250*** (8.5200) (10.4900) chE 0.0883*** 0.1310*** (6.5100) (9.3800) Operpm 0.0341*** 0.0337*** (5.1100) (6.3300) ROE 0.0657*** 0.0498*** (6.0900) (7.2000) AstTvr -0.0025 0.1940* 0.4870*** 0.0111 0.2410*** 0.3800*** 0.7890*** 0.3120*** (-0.0200) (1.9200) (4.3100) (0.1100) (2.7800) (3.9600) (7.4000) (3.3900) Lev% 0.0097*** -0.0063* -0.0016 -0.0052 0.0058** -0.0078*** -0.0028 -0.0090*** (2.7700) (-1.9400) (-0.4700) (-1.5700) (2.0300) (-2.8100) (-1.0600) (-3.3300) Size 0.0590* 0.1150*** 0.0765** 0.0666* 0.0738*** 0.1120*** 0.0789*** 0.0852*** (1.7200) (3.3500) (2.2600) (1.9600) (2.6700) (3.9200) (2.8700) (3.1300) Liq -0.0130 -0.0275 -0.0352 -0.0226 -0.0033 -0.0157 -0.0220 -0.0158 (-0.6400) (-1.4300) (-1.5800) (-1.1700) (-0.1800) (-0.9100) (-1.1800) (-0.9200) MktBk 0.0881 -0.1320* -0.0488 0.0575 0.0894* -0.0957 -0.0110 0.0438 (1.1400) (-1.7600) (-0.7400) (0.7200) (1.8200) (-1.6100) (-0.2400) (0.8900) Constant -1.9120* -1.3500 -1.4550 -1.4600 -2.9500** -0.9580 -2.5660** -2.5240** (-1.9300) (-1.3800) (-1.4500) (-1.4800) (-2.5700) (-0.3800) (-2.4300) (-2.0700) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 2939.0000 2939.0000 2939.0000 2939.0000 3699.0000 3699.0000 3699.0000 3699.0000 PseudoR2 0.1300 0.1290 0.1070 0.1240 0.1010 0.1590 0.0774 0.0845 chi2 316.8000 280.3000 272.3000 276.0000 326.5000 278.9000 251.7000 286.0000 P-value 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Panel C: DOMD vs. DNCD Variables Model 9 Model 10 Model 11 Model 12 ROA -0.1380*** (-9.2600) chE -0.0609*** (-6.0100) Operpm -0.0386*** (-7.0000) ROE -0.0635*** (-6.7500) Panel D: DOMD vs. DDCD Model 13 Model 14 Model 15 Model 16 -0.1590*** (-11.4600) -0.0313*** (-6.3600) -0.0444*** (-7.2800) -0.0729*** (-9.4900) Panel D: DOMD vs. DDCD Model 13 Model 14 Model 15 Model 16 -0.1040 -0.3840*** -0.4120*** -0.1130 (-0.9900) (-3.5500) (-3.8600) (-1.0300) 0.0127*** 0.0213*** 0.0193*** 0.0198*** (3.8200) (6.4600) (5.8700) (5.7600) -0.3060*** -0.3580*** -0.3100*** -0.3070*** (-8.3400) (-9.7300) (-8.6900) (-8.1000) 0.0011 -0.0019 -0.0082 0.0063 (0.0600) (-0.1000) (-0.3000) (0.3300) 0.0898** 0.2570*** 0.1530*** 0.1010** (1.9800) (5.0300) (3.3400) (2.2100) 6.9260*** 6.4310*** 6.5090*** 6.6760*** (6.1100) (6.1600) (5.8200) (5.6500) Yes Yes Yes Yes 2437.0000 2437.0000 2437.0000 2437.0000 0.2010 0.1310 0.1730 0.2000 344.3000 283.4000 313.5000 314.0000 0.0000 0.0000 0.0000 0.0000 Panel C: DOMD vs. DNCD Variables Model 9 Model 10 Model 11 Model 12 AstTvr -0.2480** -0.4380*** -0.5360*** -0.2890** (-2.0900) (-3.5600) (-4.3300) (-2.4100) Lev% 0.0136*** 0.0212*** 0.0218*** 0.0207*** (3.3700) (5.4100) (5.6400) (5.2300) Size -0.3570*** -0.4060*** -0.3750*** -0.3630*** (-7.9600) (-9.0400) (-8.4500) (-8.1500) Liq -0.0263 -0.0232 -0.0328 -0.0211 (-1.2000) (-1.1500) (-1.2300) (-1.0300) MktBk 0.1240** 0.2990*** 0.1850*** 0.1410** (2.0200) (4.1900) (2.9300) (2.2100) Constant 5.9000*** 5.9960*** 5.3350*** 5.8990*** (4.2400) (4.7600) (3.0700) (4.7200) Industry & Country Dummies Yes Yes Yes Yes N 1728.0000 1728.0000 1728.0000 1728.0000 PseudoR2 0.2100 0.1950 0.1840 0.2030 chi2 276.4000 237.4000 281.7000 231.9000 P-value 0.0000 0.0000 0.0000 0.0000 Source: Author’s own work. Panel E: DDCD vs. DNCD Variables Model 17 Model 18 Model 19 Model 20 ROA -0.00950 (-0.90000) chE -0.04430*** (-4.02000) Operpm 0.0014 (0.4400) ROE 0.0023 (0.4400) AstTvr -0.17900* -0.16200 -0.1960* -0.2110** (-1.72000) (-1.59000) (-1.9400) (-2.0300) Lev 0.00110 0.00140 0.0021 0.0019 (0.35000) (0.44000) (0.6700) (0.6300) Size -0.00840 -0.01090 -0.0136 -0.0138 (-0.24000) (-0.31000) (-0.3900) (-0.4000) Source: Author’s own work. Liq -0.02490 -0.02230 -0.0248 -0.0248 (-1.34000) (-1.19000) (-1.3300) (-1.3300) MktBk -0.00741 0.03070 0.0169 0.0191 (-0.12000) (0.48000) (0.2800) (0.3100) Constant 0.82200 0.55000 0.8680 0.8610 (0.84000) (0.55000) (0.9000) (0.8900) Industry & Country Dummies Yes Yes Yes Yes N 2261.00000 2261.00000 2261.0000 2261.0000 PseudoR2 0.06960 0.08680 0.0694 0.0694 chi2 181.50000 184.30000 181.0000 181.2000 P-value 0.00000 0.00000 0.0000 0.0000 Table 5 (B): Dividend Changes during COVID-19 Table 6: Dividend reductions vs. No-change dividends Panel A: (25% > DDCD) vs. DNCD Panel B: (25%≤ DDCD <50%) vs. DNCD Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 ROA 0.0685*** -0.0134 (4.7400) (-0.8700) chE -0.0083 -0.0447*** (-0.9800) (-2.6100) Operpm 0.0127* 0.0061 (1.7800) (1.4200) ROE 0.0393*** 0.0015 (5.1800) (0.2000) AstTvr -0.1190 0.0440 0.1010 -0.1230 -0.3370* -0.3410** -0.3440** -0.3790** (-0.8300) (0.3300) (0.7300) (-0.8600) (-1.9300) (-2.0200) (-2.0400) (-2.1600) Lev% 0.0081* 0.0005 0.0022 0.0003 -0.0017 -0.0014 0.0005 -0.0004 (1.7900) (0.1300) (0.5100) (0.0800) (-0.3600) (-0.3000) (0.1100) (-0.1000) Size 0.0670 0.1050** 0.0809 0.0652 -0.0301 -0.0379 -0.0475 -0.0377 (1.3100) (2.0900) (1.5700) (1.2700) (-0.5700) (-0.7200) (-0.8900) (-0.7100) Liq 0.0007 -0.0042 -0.0078 -0.0052 -0.0221 -0.0214 -0.0229 -0.0225 (0.0300) (-0.1700) (-0.2800) (-0.2000) (-0.9000) (-0.8600) (-0.8900) (-0.9200) MktBk 0.0886 -0.0642 -0.0058 0.0939 -0.0756 -0.0497 -0.0122 -0.0363 (1.0800) (-0.7000) (-0.0700) (1.1500) (-0.7100) (-0.4800) (-0.1300) (-0.3600) Constant -2.0020 -2.3200* -1.9110 -1.7160 0.1180 -0.0367 0.3920 0.2080 (-1.5500) (-1.7800) (-1.4700) (-1.3100) (0.0800) (-0.0200) (0.2800) (0.1400) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 1315.0000 1315.0000 1315.0000 1315.0000 1246.0000 1246.0000 1246.0000 1246.0000 PseudoR2 0.1020 0.0896 0.0933 0.1050 0.0816 0.0945 0.0825 0.0812 chi2 147.9000 129.0000 134.9000 151.7000 111.6000 113.9000 112.3000 110.3000 P-value 0.000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Panel C: (50% ≤ DDCD < 75%) vs. DNCD Panel D: (75% ≤ DDCD < 100%) vs. DNCD Variables Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16 ROA -0.0764*** -0.1110*** (-3.7600) (-3.6300) chE -0.0537*** -0.0706*** (-3.0500) (-3.5600) Operpm -0.0071 -0.0168* (-1.2200) (-1.6500) ROE -0.0196 -0.0238* (-1.3900) (-1.7700) Panel C: (50% ≤ DDCD < 75%) vs. DNCD Panel D: (75% ≤ DDCD < 100%) vs. DNCD Variables Model 9 Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 16 AstTvr -0.0335 -0.1090 -0.2000 -0.1010 -0.6770** -0.6250* -0.9510*** -0.8010** (-0.2100) (-0.6800) (-1.2300) (-0.6100) (-2.0600) (-1.9000) (-2.7500) (-2.2700) Lev% -0.0033 0.0009 0.0019 0.0020 0.0036 0.0052 0.0104 0.0096 (-0.6500) (0.1800) (0.3900) (0.4100) (0.4900) (0.6700) (1.4500) (1.3000) Size -0.1020* -0.1160** -0.1210** -0.1130** -0.0554 -0.0706 -0.0708 -0.0750 (-1.9100) (-2.1500) (-2.2400) (-2.1000) (-0.6800) (-0.8700) (-0.8700) (-0.9100) Liq -0.0687** -0.0589* -0.0671* -0.0687* -0.0201 -0.0200 -0.0238 -0.0263 (-2.0000) (-1.6500) (-1.9100) (-1.9400) (-0.4700) (-0.4200) (-0.5200) (-0.5900) MktBk -0.0042 0.1530* 0.1090 0.0688 -0.2610 0.0443 -0.0811 -0.0936 (-0.0400) (1.6700) (1.2500) (0.6900) (-1.4400) (0.3000) (-0.5200) (-0.6200) Constant 1.3010 0.9340 1.2920 1.4390 0.8260 -0.2670 0.3680 0.5740 (0.8700) (0.6500) (0.8900) (1.0400) (0.4400) (-0.1500) (0.2000) (0.3000) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 1208.0000 1208.0000 1208.0000 1208.0000 927.0000 927.0000 927.0000 927.0000 PseudoR2 0.0992 0.1150 0.0864 0.0890 0.1370 0.1830 0.1170 0.1170 chi2 116.5000 109.7000 110.8000 107.1000 86.1800 83.2500 75.2400 80.7300 P-value 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 Source: Author’s own work. Table 7 (A): Dividend Changes during COVID-19: Sample Sensitivity Panel A: DICD vs. DNCD Panel B: DICD vs. DDCD Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 ROA 0.13800*** 0.0788*** (5.13000) (5.2700) chE 0.08010*** 0.0757*** (3.39000) (5.5000) Operpm 0.02370*** 0.0260*** (2.71000) (2.9500) ROE 0.06730*** 0.0272*** (4.47000) (4.2600) AstTvr 0.04530 0.22600 0.53300** 0.06230 0.3650*** 0.5020*** 0.8160*** 0.4400*** (0.24000) (1.21000) (2.47000) (0.31000) (2.8200) (3.7500) (5.3700) (3.3500) Lev% 0.01280** -0.00241 -0.00121 -0.00315 -0.0049 -0.0120*** -0.0104** -0.0154*** (1.97000) (-0.40000) (-0.21000) (-0.53000) (-1.1700) (-2.8000) (-2.5200) (-3.7600) Size -0.05470 -0.02690 -0.05890 -0.05220 0.05280 0.0649 0.0538 0.0547 (-0.84000) (-0.40000) (-0.93000) (-0.81000) (1.2100) (1.4400) (1.2100) (1.2600) Liq -0.00687 -0.01920 -0.02820 -0.02290 -0.0175 -0.0307 -0.0133 -0.0289 (-0.22000) (-0.63000) (-0.95000) (-0.76000) (-0.7000) (-1.1600) (-0.5200) (-1.1700) MktBk 0.05320 -0.27400* -0.14000 0.00248 0.1140* -0.0210 0.0645 0.0741 (0.44000) (-1.86000) (-1.15000) (0.02000) (1.9200) (-0.3100) (1.1300) (1.3200) Constant 0.69200 1.92000 1.80500 1.36300 -1.6890 -0.2990 -1.6670 -1.1770 (0.44000) (1.25000) (1.18000) (0.89000) (-1.3100) (-0.1600) (-1.2900) (-0.9400) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 819.00000 819.00000 819.00000 819.00000 1303.0000 1303.0000 1303.0000 1303.0000 PseudoR2 0.17200 0.16900 0.13700 0.17100 0.1260 0.1700 0.1200 0.1160 chi2 123.80000 96.31000 93.53000 116.70000 162.4000 159.8000 146.7000 161.3000 P-value 0.00000 0.00000 0.00000 0.00000 0.0000 0.0000 0.0000 0.0000 Panel C: DOMD vs. DNCD Variables Model 9 Model 10 Model 11 Model 12 ROA -0.14800*** (-5.53000) chE -0.0569*** (-2.7500) Operpm -0.0349*** (-3.5400) ROE -0.0611*** (-2.9800) Panel D: DOMD vs. DDCD Model 13 Model 14 Model 15 Model 16 -0.1720*** (-6.7200) -0.0251*** (-3.5400) -0.0388*** (-4.5600) -0.0688*** (-5.5300) Panel C: DOMD vs. DNCD Model 9 Model 10 Model 11 Model 12 AstTvr -0.44600** -0.5180** -0.6020*** -0.4510** (-2.10000) (-2.4500) (-2.8200) (-2.2000) Lev% 0.02330*** 0.0294*** 0.0312*** 0.0310*** (2.96000) (3.9100) (4.1500) (4.0000) Size -0.44400*** -0.4720*** -0.4460*** -0.4400*** (-5.15000) (-5.5600) (-5.4900) (-5.3500) Liq 0.00579 0.0195 0.0050 0.0197 (0.19000) (0.7000) (0.1500) (0.7100) MktBk 0.00840 0.2840** 0.1310 0.0563 (0.07000) (2.1400) (1.0500) (0.4000) Constant 7.31100*** 6.8880*** 6.3830*** 7.0830*** (3.32000) (3.5100) (2.7700) (3.6700) Industry & Country Dummies Yes Yes Yes Yes N 518.00000 518.0000 518.0000 518.0000 PseudoR2 0.28300 0.2520 0.2480 0.2620 chi2 123.60000 99.4500 121.5000 101.7000 P-value 0.00000 0.0000 0.0000 0.0000 Panel D: DOMD vs. DDCD Model 13 Model 14 Model 15 Model 16 -0.1320 -0.4490** -0.4350** -0.0968 (-0.7600) (-2.4000) (-2.3700) (-0.5500) 0.0129** 0.0203*** 0.0205*** 0.0207*** (2.4700) (3.9400) (4.0000) (3.8100) -0.2620*** -0.2980*** -0.2650*** -0.2630*** (-4.4600) (-5.0700) (-4.7200) (-4.2700) 0.0197 0.0238 0.0017 0.0332 (0.7800) (1.0200) (0.0500) (1.4100) 0.0851 0.2570*** 0.1680** 0.1010 (1.2800) (3.3900) (2.5100) (1.4600) 6.3950*** 5.5720*** 5.6540*** 5.9140*** (4.1200) (3.9400) (3.9300) (3.7900) Yes Yes Yes Yes 951.0000 951.0000 951.0000 951.0000 0.2390 0.1480 0.1910 0.2310 144.6000 124.6000 139.7000 137.7000 0.0000 0.0000 0.0000 0.0000 Source: Author’s own work. Source: Author’s own work. Panel E: DDCD vs. DNCD Variables Model 17 Model 18 Model 19 Model 20 ROA 0.00845 (0.47000) chE -0.03980** (-2.29000) Operpm 0.01190* (1.85000) ROE 0.000199 (0.030000) AstTvr -0.24500 -0.25300 -0.29300 -0.223000 (-1.23000) (-1.24000) (-1.48000) (-1.130000) Lev% 0.01060* 0.00703 0.01000* 0.009840* (1.84000) (1.22000) (1.83000) (1.770000) Size -0.06630 -0.06790 -0.07340 -0.064800 (-1.03000) (-1.04000) (-1.14000) (-1.010000) Liq -0.00905 -0.01230 -0.00991 -0.009840 (-0.32000) (-0.44000) (-0.35000) (-0.340000) MktBk -0.17200 -0.17600 -0.14100 -0.194000 (-1.35000) (-1.51000) (-1.16000) (-1.560000) Constant 1.76400 1.76900 1.88400 1.791000 (1.16000) (1.13000) (1.24000) (1.180000) Industry & Country Dummies Yes Yes Yes Yes N 951.00000 951.00000 951.00000 951.000000 PseudoR2 0.10800 0.12400 0.11100 0.108000 chi2 78.96000 78.61000 81.06000 79.100000 P-value 0.00000 0.00000 0.00000 0.000000 Table 7 (B): Dividend Changes during COVID-19: Sample Sensitivity Table 8: Dividend Increases vs. No-change dividends Variables Panel A: (25% > DICD) vs. DNCD Panel B: (25% ≤ DICD < 50%) vs. DNCD ROA 0.1350*** 0.1540*** (8.2500) (7.7500) chE 0.0604*** 0.1250*** (4.3900) (5.4600) Operpm 0.0414*** 0.05710*** (6.0400) (6.50000) ROE 0.0708*** 0.08600*** (7.4300) (7.36000) AstTvr 0.0845 0.3560*** 0.6580*** 0.1000 0.1160 0.3830*** 0.75400*** 0.13700 (0.6900) (2.9300) (5.0500) (0.8000) (0.7100) (2.6100) (5.04000) (0.83000) Lev% 0.0095** -0.0064 -0.0015 -0.0066* 0.0104** -0.0104** -0.00080 -0.00890* (2.2600) (-1.6300) (-0.3800) (-1.6600) (1.9900) (-2.1100) (-0.17000) (-1.84000) Size 0.1250*** 0.1980*** 0.1460*** 0.1340*** -0.0060 0.0584 0.00121 -0.00202 (2.8300) (4.5300) (3.3400) (3.0300) (-0.1100) (1.0400) (0.02000) (-0.04000) Liq -0.0064 -0.0110 -0.0348 -0.0151 -0.0555* -0.0705** -0.10500*** -0.06600** (-0.2500) (-0.4800) (-1.2100) (-0.6300) (-1.7500) (-2.3200) (-3.13000) (-2.11000) MktBk 0.0722 -0.2120** -0.0702 0.0568 0.1090 -0.1700 -0.01400 0.11700 (0.7100) (-2.1400) (-0.7600) (0.5800) (0.9300) (-1.5800) (-0.14000) (1.03000) Constant -4.2800*** -4.2410*** -4.0570*** -3.8980*** -2.7870** -2.1180* -1.97000 -1.96100 (-3.2600) (-3.5400) (-3.2200) (-2.8800) (-2.0700) (-1.6600) (-1.56000) (-1.46000) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 1544.0000 1544.0000 1544.0000 1544.0000 1230.0000 1230.0000 1230.00000 1230.00000 PseudoR2 0.1300 0.1020 0.1120 0.1270 0.1620 0.1680 0.14100 0.16100 chi2 209.0000 166.7000 180.4000 202.1000 178.1000 147.4000 168.90000 173.70000 P-value 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.00000 0.00000 Variables Panel C: (50% ≤ DICD < 75%) vs. DNCD Panel D: (75% ≤ DICD <100%) vs. DNCD ROA 0.1300*** 0.1650*** (6.7900) (5.3900) chE 0.138*** 0.1470*** (5.200) (3.3500) Operpm 0.032*** 0.0446*** (3.700) (3.6300) ROE 0.0739*** 0.0983*** (6.6400) (5.5900) AstTvr -0.0209 0.228 0.463*** -0.0065 0.4740* 0.7320*** 0.9760*** 0.5010* (-0.1200) (1.310) (2.790) (-0.0400) (1.7600) (2.9000) (4.2700) (1.8100) Variables Panel C: (50% ≤ DICD < 75%) vs. DNCD Panel D: (75% ≤ DICD <100%) vs. DNCD Lev% 0.0039 -0.017*** -0.007 -0.0121** 0.0113 -0.0158 -0.0012 -0.0107 (0.6700) (-3.050) (-1.340) (-2.2400) (1.0900) (-1.5100) (-0.1200) (-1.0700) Size 0.0178 0.111 0.042 0.0194 0.0828 0.1800 0.0639 0.0725 (0.2600) (1.620) (0.620) (0.2900) (0.6100) (1.2000) (0.4500) (0.5200) Liq -0.0044 -0.025 -0.032 -0.0124 0.0088 -0.0215 -0.0207 -0.0167 (-0.1400) (-0.870) (-0.910) (-0.4200) (0.1900) (-0.4000) (-0.3300) (-0.3400) MktBk 0.0524 -0.189 -0.132 0.0510 0.3600 0.2000 0.2290 0.3620* (0.4800) (-1.400) (-1.030) (0.4700) (1.5800) (0.6200) (0.8300) (1.7300) Constant -2.2850 -2.000 -2.150 -1.6980 -6.0130** -6.1080** -4.8670* -4.8860* (-1.2800) (-1.060) (-1.160) (-0.9500) (-2.4100) (-2.2700) (-1.9500) (-1.9100) Industry & Country Dummies Yes Yes Yes Yes Yes Yes Yes Yes N 1093.0000 1093.000 1093.000 1093.0000 848.0000 848.0000 848.0000 848.0000 PseudoR2 0.1230 0.155 0.097 0.1220 0.2000 0.2300 0.1630 0.2060 chi2 125.1000 106.000 100.700 122.1000 95.1200 82.8800 91.2900 99.7700 P-value 0.0000 0.000 0.000 0.0000 0.0000 0.0000 0.0000 0.0000 Source: Author’s own work.