Corporate payout, cash holdings, and the COVID-19 crisis: Evidence from the G-7 countries.

This paper examines the impact of the COVID-19 pandemic on the adjustments of dividends and share repurchases of publicly listed firms in the G-7 countries. Firms in the United Kingdom, Germany, France, and Italy experienced a widespread cut in dividends, while firms in the United States and Canada cut cash payout more via share repurchases, with Japanese firms in between. Corporate cash holdings helped mitigate the negative impact of COVID on payout adjustments, but the impact was less significant for European firms.

Introduction

COVID-19, declared by the World Health Organization (WHO) as a global pandemic in March 2020, is a health crisis that caused tremendous economic disruption around the world. For many businesses, it led to a significant drop in their cash flows and induced higher uncertainty about the future business perspective. Hence, surviving the pandemic became a top priority for corporate executives, especially during the early stage of the COVID crisis. As Iyer and Simkins (2022, p. 13) point out, “The corporate response to the COVID crisis with respect to survival mode financing, investing, and payout policies is still emerging across the world. More research is needed in this area”. In this study, we aim to bring new insights on the business impacts of COVID by exploring corporate adjustments of payout policies, including both dividends and share repurchases, in an international setting, with emphasis on potential differences across countries.

Cash dividends and share repurchase are the two channels employed by corporations to distribute cash to their shareholders. Corporations are very conservative on dividend decisions because maintaining historical dividend levels is regarded as a strong commitment to shareholders, and dividend cuts are only considered in extreme situations. Share repurchases tend to be more flexible and are often considered a substitute for dividends (DeAngelo and DeAngelo, 1990; Grullon and Michaely, 2002; Brav et al., 2005). In recent years, firms in the U.S. have distributed more cash via share repurchases than dividends, and in many other countries share repurchases have gradually grown after regulatory restrictions were eased (Lee and Suh, 2011; Floyd et al., 2015; Wang et al., 2021). Hence, a comprehensive analysis of COVID impacts on payout policy requires us to examine both dividends and share repurchases.

Further, we should be aware of potential differences in payout adjustments across countries. As reviewed by Booth and Zhou (2017), past studies have found that legal protection, financial market development, economic growth, politics, and culture contribute to differences in corporate payout policies across countries. During the COVID crisis, governments differed in their responses, and some may have impact on corporate cash payouts. For example, the Coronavirus Large Business Interruption Loan Scheme (CLBILS) provided by the UK government required that “companies borrowing more than £50 million through CLBILS will be subject to restrictions on dividend payments, senior pay and share buy-backs during the period of the loan”.1 However, the Canada Emergency Wage Subsidy (CEWS) program provided by the federal government imposed no restrictions on corporate payout.2 Thus, studying payout adjustments in an international setting may offer new insights.

Using a sample of listed firms from G-7 countries, we find that many firms reduced cash payout through dividends and share repurchases during the COVID year of 2020, compared to pre-COVID years. However, the scales of adjustment through these two channels vary across countries. Firms in Europe (UK, Germany, France, and Italy) experienced a widespread cut in dividends, while firms in the United States and Canada cut cash payout more via share repurchases, with Japanese firms in between. Furthermore, we use logistic regressions to quantify COVID's impact on these adjustments, and explore whether corporate cash holdings provided a mitigating effect. Due to costly external financing, firms have incentives to hoard cash to mitigate the negative impacts of unexpected shocks to cash flows or investment opportunities (Opler et al., 1999; Bates et al., 2009). This precautionary cash holding—an important component of financial flexibility—was particularly important during the onset of COVID. Several recent studies on COVID show that firms with more cash reserves suffered milder declines in stock returns (Fahlenbrach et al., 2021; Ding et al., 2021), had better operating performance, and experienced less several impact on investments (Zheng, 2022; Tawiah and Keefe, 2022), compared with cash-poor peers. Corporate payout policy is about distributing cash to shareholders. When COVID hit, firms with more cash in hand should face less pressure to preserve and enhance financial flexibility by cutting cash payout, or have more financial capacity to maintain or increase payout. Hence, cash holdings should help mitigate the negative impact of COVID shock on corporate payouts. This prediction is supported by the findings of our empirical tests. We show that firms are more (less) likely to decrease (increase) their dividends or share repurchases after COVID hit, but COVID's impact is less severe for firms with more cash holdings. The mitigating effect of cash holdings is strong in North America and Japan, but lacks statistical significance for European firms.

Our paper adds to several strands of literature. First, we contribute to the emerging literature examining COVID-19 impacts on corporate decisions, such as corporate investments, financing, and payout policies (Krieger et al., 2021; Zheng, 2022; Tawiah and Keefe, 2022; Ali, 2022; Pettenuzzo et al., 2021). In particular, the pioneering work by Krieger et al. (2021) and Ali (2022) study the adjustment of dividend policy in the U.S. and a pooled sample of G-13 countries, and Pettenuzzo et al. (2021) examine reductions in dividends and share repurchases in the U.S. We add new insights by examining changes in both dividends and share repurchases, and documenting the differences and similarities across countries. Second, we add to the literature exploring the importance of financial flexibility on corporate decisions, particularly during challenging situations caused by unexpected exogenous shocks. Bliss et al. (2015) study U.S. firms’ payouts during the 2008–09 financial crisis and find that payout reductions are more likely in firms with less financial flexibility. Zheng (2022) and Tawiah and Keefe (2022) show the cash holdings help mitigate the negative impact of COVID on operating performance and investments in the U.S. Our paper shows that this mitigating effect also applies to corporate payout policy in an international setting, but the strength of this effect is not consistent across all countries. We believe that payout restrictions tied to COVID relief programs provided by governments may have created the cross-country differences documented in our paper, and we call for more research in the future.

Data

The initial sample used in this study includes firms listed on the main stock exchanges of the G-7 countries (Japan, Germany, France, Italy, United Kingdom, Canada, and the United States) with fiscal years ending between April 2015 and March 2021. WHO declared COVID-19 a global pandemic on March 11, 2020. To match this timeline closely, we define a fiscal year t as the 12-month range from April in year t to March in year t + 1. Hence, 2020 fiscal year in our analysis, classified as the COVID-year, includes firm-year observations with fiscal years ending between April 2020 and March 2021.3

The accounting and market data are collected from the Refinitiv Worldscope and Datastream databases. Following the literature, we remove financial firms (SIC 6000–6999) and utility firms (SIC 4900–4949). Firms that did not pay dividends or repurchase shares during our sample period are removed. Further, firms with fewer than three years of observations are removed. We exclude observations with missing information for total assets, market capitalization, as well as those with missing information on both share repurchase and dividends. Our final sample contains 37,885 observations of 6697 firms. Table 1 provides a summary of our sample distribution by year (Panel A), country (Panel B), and industry (Panel C), as well as the summary statistics of variables used in our analysis (Panel D). Japan (40.27%) and the United States (30.88%) have the largest representations, while four European countries jointly account for 23.23% of our sample.

Table 1

Sample breakdown and summary statistics.

This table presents the sample distribution by year (Panel A), country (Panel B), and industry (Panel C), as well as the summary statistics of variables used in our analysis (Panel D).

Panel A
Year	N	%
2015	6087	16.07
2016	6295	16.62
2017	6493	17.14
2018	6512	17.19
2019	6325	16.70
2020	6173	16.29
Total	37,885	100
Panel B
Country	N	%
Canada	2130	5.62
France	2339	6.17
United Kingdom (UK)	3658	9.66
Germany	1711	4.52
Italy	1091	2.88
Japan	15,256	40.27
United States (USA)	11,700	30.88
Total	37,885	100

Empirical results

Descriptive statistics

Our sample includes firms that distributed cash to shareholders via dividends or share repurchases during 2015–2020. Table 2 Panel A shows an overall larger portion of dividend payers than repurchasing firms in the G-7 countries, with the proportions of both dividend-paying firms and repurchasing firms dropping after the COVID pandemic started. Panel B reports the statistics across countries: USA is the only country with more firms repurchasing shares than paying dividends, while Japan has the highest proportion of dividend payers. More importantly, four European countries (UK, Germany, France, Italy) experienced large drops in dividend payers in 2020, while the other three countries experienced very small drops.

Table 2

Annual percentage of dividend-paying firms and repurchasing firms.

This table presents the number and percentage of firms that pay dividends or repurchase shares each year among all G-7 countries (Panel A) and in each country (Panel B). The sample contains observations of firms that had ever distributed cash payout through dividends or repurchases during the sample period.

Panel A: All G-7 countries
		Dividend-paying Firms		Repurchasing firms
Year	N	N	%	N	%
2015	6087	4412	72.5%	3135	51.5%
2016	6295	4495	71.4%	3185	50.6%
2017	6493	4611	71.0%	3255	50.1%
2018	6512	4619	70.9%	3391	52.1%
2019	6325	4530	71.6%	3353	53.0%
2020	6173	4083	66.1%	3129	50.7%

Table 3 reports aggregate amounts of dividends, repurchases, and total cash payout for each year. All values are converted to 2020 U.S. dollars for easier comparison. The aggregate amounts in all G-7 countries, as reported in Panel A, show a higher portion of cash being distributed via dividends in most years. Further, COVID caused drops in both dividends (about $85 billion) and share repurchases (about $199 billion) in 2020. Panel B shows that large variations exist among countries. U.S. firms spent much more cash repurchasing shares than paying dividends, but in European countries repurchases accounted for less than a quarter of aggregate payout amount. More importantly, the impact of COVID on total amount of dividends and share repurchases varies across countries. In USA, Canada, and Japan, the aggregate amount of share repurchases dropped much more than dividend payout. Compared to 2019 amounts, American firms have cut dividends (repurchases) by about $13.73 ($147.85) billion, Canadian firms cut dividends (repurchases) by $1.57 ($9.26) billion, and Japanese firms cut dividends (repurchases) by $16.09 ($21.60) billion. In contrast, European countries cut more dividends more than share repurchases. For example, the reduction of dividends (repurchases) in 2020 relative to 2019 is $16.41 ($2.87) billion in France, $8.43 ($1.69) billion in Italy, and $22.45 ($17.11) billion in the UK. In Germany, aggregate dividend amount dropped by $6.44 billion but aggregate repurchases increased by $1.15 billion.

Table 3

Annual aggregate dividends, repurchases, and total payout.

This table presents aggregate dividends, repurchases, and total payout each year for all G-7 countries (Panel A) and for each country (Panel B). All the amounts have been converted into 2020 U.S. dollars. The last two columns of each table report the portions of aggregate cash payout distributed through dividends and through share repurchases each year.

Panel A: All G-7 countries
		In Millions USD (2020 constant)			As the percentage of Aggregate Total Payout
	Year	Aggregate Total Payout	Aggregate Dividends	Aggregate Repurchase	Aggregate Dividends	Aggregate Repurchase
All G-7 Countries	2015	1261,405	623,580	637,825	49.4%	50.6%
	2016	1229,273	626,920	602,353	51.0%	49.0%
	2017	1216,268	679,696	536,573	55.9%	44.1%
	2018	1551,743	724,535	827,208	46.7%	53.3%
	2019	1483,019	757,655	725,363	51.1%	48.9%
	2020	1198,664	672,548	526,117	56.1%	43.9%

In Table 4 we examine payout changes by reporting the percentage of firms that adjusted cash payouts, including dividend decrease, dividend omission, dividend increase, repurchase decrease, and repurchase increase, before and after COVID. To make it easier to compare to the changes during the COVID year of 2020, we average the annual percentages during the pre-COVID period of 2015–2019. In the pooled sample of all G-7 countries, there are more dividend reductions and fewer dividend increases after COVID hit, which is similar to the finding of Ali (2022) for the G-13 countries. Meanwhile, we notice COVID had a similar impact on share repurchases, i.e. relatively more firms reducing repurchases and fewer firms increasing repurchases after COVID, but the impact on repurchases is smaller than on dividends. More interesting observations come from checking the changes in each country. Compared to other countries, the four European countries have experienced the largest increase in the percentage of firms cutting their dividends and the largest drop in the percentage of firms increasing dividends, but much smaller adjustments in repurchases. About 45% (in Germany) to 56% (in UK) of firms cut their dividends after COVID hit, compared with about 14% to 17% of firms doing so on average during pre-COVID years. The percent of firms omitting dividends also jumped from single digit (3.1% to 5.4%) to double-digit (16.6% to 32.1%). In contrast, more firms in the USA and Canada reduce share repurchases than dividends, and the difference increases after COVID hit in 2020. In that year, 43.4% of American firms cut share repurchase and 18.2% cut dividends, while 37.3% of Canadian firms cut share repurchases and 26.5% cut dividends. In Japan, the adjustments of dividends and repurchases triggered by COVID relative to earlier years are more moderate, compared with companies in North America and Europe.

Table 4

Decreases and increases in dividends and share repurchases.

This table presents the percentage of firms that adjust their dividends (decrease, omission, and increase) or share repurchases (decrease and increase) during pre-COVID years of 2015–2019 and during the COVID year of 2020. For pre-Covid years, we calculate the annual percentage of a given adjustment, and then calculate the time series average across the years 2015–2019.

		Dividends			Repurchases
		Decrease	Omission	Increase	Decrease	Increase
All G-7 Countries	Pre-COVID Years	14.6%	1.9%	45.5%	29.4%	30.5%
	COVID Year	33.5%	7.3%	28.0%	36.6%	25.8%
	Difference	18.9%	5.5%	-17.5%	7.2%	-4.7%
USA	Pre-COVID Years	10.4%	1.8%	33.3%	33.6%	36.8%
	COVID Year	18.2%	2.1%	25.0%	43.4%	29.8%
	Difference	7.8%	0.4%	-8.4%	9.8%	-7.1%
Canada	Pre-COVID Years	15.5%	2.5%	38.3%	21.3%	26.9%
	COVID Year	26.5%	3.0%	29.5%	37.3%	22.6%
	Difference	11.0%	0.5%	-8.8%	16.1%	-4.3%
UK	Pre-COVID Years	14.4%	3.3%	60.6%	20.7%	21.7%
	COVID Year	55.8%	27.3%	22.3%	26.9%	18.9%
	Difference	41.4%	23.9%	-38.3%	6.2%	-2.8%
Germany	Pre-COVID Years	17.3%	4.1%	49.5%	11.7%	11.0%
	COVID Year	45.0%	16.6%	20.1%	11.8%	10.7%
	Difference	27.7%	12.5%	-29.4%	0.1%	-0.2%
France	Pre-COVID Years	17.0%	3.1%	38.8%	29.0%	27.8%
	COVID Year	48.0%	24.4%	11.9%	30.9%	27.1%
	Difference	30.9%	21.3%	-26.8%	1.9%	-0.7%
Italy	Pre-COVID Years	17.3%	5.4%	44.3%	19.6%	23.3%
	COVID Year	46.2%	32.1%	20.7%	27.2%	23.4%
	Difference	28.9%	26.7%	-23.7%	7.5%	0.1%
Japan	Pre-COVID Years	16.9%	0.8%	53.0%	32.2%	31.4%
	COVID Year	36.4%	2.0%	35.1%	37.9%	26.6%
	Difference	19.5%	1.2%	-17.9%	5.6%	-4.8%

Overall, these descriptive statistics show that COVID has caused firms to cut cash payout in 2020, but the relative impact on dividends and share repurchases varies across countries.4

Regression results

Following the research design of Bliss et al. (2015), Tawiah and Keefe (2022), and Zheng (2022), we formally estimate the impact of COVID-19 on corporate decisions to adjust dividends or share repurchases, as well as the role of corporate cash holdings, by estimating the following logistic regression model.We consider four types of payout changes, i.e. dividend decrease, repurchase decrease, dividend increase, and repurchase increase, respectively. For each type, the dependent variable, Payout Change, is set to 1 if firm i makes the change in year t, and 0 otherwise. COVID is an indicator variable, equal to 1 for fiscal year ends between April 2020 and March 2021, the 12-month period after COVID-19 was declared a global pandemic, and 0 otherwise. The impact of COVID on a given type of payout change is measured by β1. The mitigating effect of cash holdings during COVID year is measured by β2. Control variables, X, include leverage (TD/TA), stock return volatility (Volatility), profitability, size, retained earnings (RE/TA), and sales growth rate (SGR). Variable definitions are presented in Appendix 2. Industry fixed effects and country fixed effects are included, and standard errors are clustered at firm level. Besides running regressions for the pooled sample of all G-7 countries, we consider differences across geographic regions as identified in the previous section by running separate regressions for three subgroups: North America (USA and Canada), Europe (UK, Germany, France, and Italy), and Japan.

Table 5 presents the results of our regression model with Dividend Decrease as the dependent variable, and here the sample includes firms with positive dividend payout in the prior year. The coefficient estimates on COVID are all positive and statistically significant, indicating that firms were more inclined to cut their dividends after COVID hit. The coefficient estimate of the interaction term is negative, indicating that cash holdings help firms mitigate the negative impact of COVID. Although the mitigation effect is strong for North America and Japan, it lacks statistical significance for European firms. We further examine the economic significance of COVID's impact by calculating the change in the probability of dividend decrease due to COVID for an average firm in each region; we report the results in Panel B. The largest impact is observed for Europe where COVID increased the probability of dividend decrease by 47.3 percentage points, while the smallest impact is observed for North America (13.2 percentage points). These are consistent with the descriptive statistics in Table 4 of much higher percentages of dividend decreases in the UK, Germany, France, and Italy, compared with the USA and Canada. In Table 6 , the dependent variable is the reduction in share repurchases, and here the sample only includes firms with positive share repurchase in the prior year. The positive coefficient estimate of COVID indicates that firms are more inclined to reduce share repurchase after COVID, and coefficient estimates vary across regions. Moreover, the mitigation effect of cash holdings is strong for firms in North America and Japan, but lacks statistical significance for European firms. In terms of economic significance, as measured by the change in probability of repurchase decrease due to COVID, the impact of COVID is smaller for firms in Europe than for those in North America and Japan.

Table 5

Logit regression of reduction in dividends.

This table reports the coefficient estimates of logistic regressions used to explain a firm's decision to decrease its dividends. Column [1] reports the results for the pooled sample of firms from all G-7 countries, while Columns [2]–[4] are for subsamples of firms from Europe (UK, Italy, Germany, France), North America (Canada and USA), and Japan, respectively. The dependent variable is Dividend Decrease, equal to 1 if a firm's dividend amount decreases 1% or more from the previous year's dividend amount, and 0 otherwise. The sample includes firms with positive dividend payout in the prior year. Industry fixed effects are included in all the regressions. Country fixed effects are included except for the regression for Japan. Standard errors are clustered at firm level. Robust standard errors are reported in brackets. *** p<0.01, ** p<0.05, * p<0.1. Panel B reports the change in probability of dividend decrease due to COVID for an average firm in each sample. It is estimated by using the coefficient estimates reported in Panel A and the mean values of explanatory variables in the corresponding sample.

Panel A:	Dividends Decrease
	All G-7 Countries	Europe	North America	Japan
	[1]	[2]	[3]	[4]
COVID	1.440***	2.203***	0.872***	1.297***
	[0.057]	[0.125]	[0.100]	[0.089]
COVID*Cash/TA	-1.404***	-0.924	-1.359**	-1.500***
	[0.254]	[0.679]	[0.596]	[0.347]
Cash/TA	1.030***	0.348	0.907***	1.446***
	[0.168]	[0.355]	[0.349]	[0.229]
TD/TA	0.613***	0.575**	1.013***	0.712***
	[0.133]	[0.265]	[0.236]	[0.197]
Volatility	2.500***	6.141***	8.581***	-0.042
	[0.359]	[0.847]	[0.913]	[0.503]
Profitability	-4.725***	-4.214***	-2.941***	-6.608***
	[0.381]	[0.554]	[0.595]	[0.952]
Size	-0.085***	-0.091***	-0.150***	-0.044***
	[0.011]	[0.019]	[0.026]	[0.017]
RE/TA	-0.396***	-0.245*	-0.353***	0.054
	[0.093]	[0.139]	[0.109]	[0.181]
SGR	-1.651***	-1.657***	-0.933***	-2.872***
	[0.179]	[0.349]	[0.188]	[0.395]
Constant	-0.162	-0.024	0.050	-0.971***
	[0.184]	[0.314]	[0.386]	[0.258]
Industry FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	No
Observations	25,767	6111	6380	13,276
Pseudo R-squared	0.097	0.164	0.108	0.088
Panel B:
Change in the probability of dividend decrease due to COVID	+24.8%	+47.3%	+13.2%	+17.0%

Table 6

Logit regression of reduction in share repurchases.

This table reports the coefficient estimates of logistic regressions used to explain a firm's decision to decrease its share repurchases. Column [1] reports the results for the pooled sample of firms from all G-7 countries, while Columns [2]–[4] are for subsamples of firms from Europe (UK, Italy, Germany, France), North America (Canada and USA), and Japan, respectively. The dependent variable is Repurchase Decrease, equal to 1 if a firm's share repurchase amount decreases 1% or more from the previous year's repurchase amount, and 0 otherwise. The sample includes firms with positive share repurchases in the prior year. Industry fixed effects are included in all the regressions. Country fixed effects are included except for the regression for Japan. Standard errors are clustered at firm level. Robust standard errors are reported in brackets. *** p<0.01, ** p<0.05, * p<0.1. Panel B reports the change in probability of repurchase decrease due to COVID for an average firm in each sample. It is estimated by using the coefficient estimates reported in Panel A and the mean values of explanatory variables in the corresponding sample.

Panel A:	Repurchase Decrease
	All G-7 Countries	Europe	North America	Japan
	[1]	[2]	[3]	[4]
COVID	0.561***	0.334*	0.663***	0.543***
	[0.063]	[0.172]	[0.088]	[0.108]
COVID*Cash/TA	-0.863***	-0.554	-0.962**	-0.897**
	[0.237]	[0.855]	[0.385]	[0.355]
Cash/TA	-0.156	-0.770***	-0.159	-0.078
	[0.109]	[0.298]	[0.158]	[0.199]
TD/TA	0.541***	0.114	0.505***	0.781***
	[0.087]	[0.274]	[0.121]	[0.168]
Volatility	1.583***	2.644**	0.898	1.739***
	[0.343]	[1.137]	[0.561]	[0.512]
Profitability	-0.772***	-0.816**	-0.959***	-0.881**
	[0.136]	[0.333]	[0.187]	[0.367]
Size	-0.084***	-0.119***	-0.064***	-0.085***
	[0.008]	[0.020]	[0.014]	[0.013]
RE/TA	0.031	-0.036	0.004	0.260**
	[0.020]	[0.085]	[0.023]	[0.109]
SGR	-0.440***	-0.268*	-0.611***	-0.113
	[0.068]	[0.140]	[0.096]	[0.159]
Constant	1.171***	2.169***	1.168***	1.029***
	[0.139]	[0.351]	[0.231]	[0.195]
Industry FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	No
Observations	18,402	2729	7558	8115
Pseudo R-squared	0.019	0.029	0.023	0.013
Panel B:
Change in the probability of repurchase decrease due to COVID	+9.6%	+5.4%	+11.5%	+8.2%

The statistics reported in Table 4 show that every year some firms increased dividends or repurchases, but the percentage of those firms has dropped after COVID hit. In Tables 7 and 8 , we formally test the impact of COVID on the decision to increase cash payout via dividends and repurchases, respectively. The results reported in Table 7 (Table 8) show that firms are less likely to increase their dividends (share repurchases) after COVID hit, as shown by the negative coefficient estimates for COVID indicator variable. The positive coefficient estimates of COVID*Cash/TA interaction term indicate that the negative impact of COVID on payout increases is less severe for firms with more cash holdings. For European firms, the negative impact of COVID on dividend increase is much greater than other regions, while the negative impact on share repurchase increase is much smaller than other regions and lacks statistical significance, and the mitigation effect of cash holdings lacks statistical significance.

Table 7

Logit regression of increase in dividends.

This table reports the coefficient estimates of logistic regressions used to explain a firm's decision to increase its dividends. Column [1] reports the results for the pooled sample of firms from all G-7 countries, while Columns [2]–[4] are for subsamples of firms from Europe (UK, Italy, Germany, France), North America (Canada and USA), and Japan, respectively. The dependent variable is Dividend Increase, equal to 1 if a firm's dividend amount increases 1% or more from the previous year's dividend amount, and 0 otherwise. Industry fixed effects are included in all the regressions. Country fixed effects are included except for the regression for Japan. Standard errors are clustered at firm level. Robust standard errors are reported in brackets. *** p<0.01, ** p<0.05, * p<0.1. Panel B reports the change in probability of dividend increase due to COVID for an average firm in each sample. It is estimated by using the coefficient estimates reported in Panel A and the mean values of explanatory variables in the corresponding sample.

Panel A:	Dividends Increase
	All G-7 Countries	Europe	North America	Japan
	[1]	[2]	[3]	[4]
COVID	-1.035***	-1.793***	-0.411***	-1.066***
	[0.049]	[0.122]	[0.071]	[0.081]
COVID*Cash/TA	0.911***	0.476	0.800**	1.029***
	[0.202]	[0.646]	[0.396]	[0.280]
Cash/TA	-0.982***	-0.698**	-0.547**	-1.778***
	[0.149]	[0.324]	[0.278]	[0.218]
TD/TA	-0.441***	-0.398	-0.427**	-0.692***
	[0.123]	[0.256]	[0.209]	[0.188]
Volatility	-2.095***	-5.127***	-11.442***	2.338***
	[0.322]	[0.735]	[0.866]	[0.425]
Profitability	6.306***	6.759***	3.924***	8.289***
	[0.294]	[0.481]	[0.422]	[0.756]
Size	0.201***	0.159***	0.231***	0.212***
	[0.010]	[0.018]	[0.023]	[0.015]
RE/TA	0.656***	0.683***	0.419***	0.784***
	[0.081]	[0.121]	[0.101]	[0.147]
SGR	0.455***	0.725***	0.177*	0.979***
	[0.070]	[0.141]	[0.101]	[0.217]
Constant	-3.715***	-2.199***	-2.775***	-3.353***
	[0.176]	[0.283]	[0.362]	[0.233]
Industry FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	No
Observations	35,357	8196	12,541	14,620
Pseudo R-squared	0.154	0.206	0.202	0.110
Panel B:
Change in the probability of dividend increase due to COVID	-12.0%	-30.4%	-4.7%	-19.4%

Table 8

Logit regression of increase in share repurchases.

This table reports the coefficient estimates of logistic regressions used to explain a firm's decision to increase repurchases. Column [1] reports the results for the pooled sample of firms from all G-7 countries, while Columns [2]–[4] are for the subsamples of firms from Europe (UK, Italy, Germany, France), North America (Canada and USA), and Japan, respectively. The dependent variable is Repurchase Increase, equal to 1 if a firm's share repurchase amount increases 1% or more from the previous year's repurchase amount, and 0 otherwise. Industry fixed effects are included in all the regressions. Country fixed effects are included except for the regression for Japan. Standard errors are clustered at firm level. Robust standard errors are reported in brackets. *** p<0.01, ** p<0.05, * p<0.1. Panel B reports the change in probability of repurchase increase due to COVID for an average firm in each sample. It is estimated by using the coefficient estimates reported in Panel A and the mean values of explanatory variables in the corresponding sample.

Panel A:	Repurchase Increase
	All G-7 Countries	Europe	North America	Japan
	[1]	[2]	[3]	[4]
COVID	-0.398***	-0.168	-0.456***	-0.476***
	[0.049]	[0.109]	[0.071]	[0.085]
COVID*Cash/TA	0.818***	0.543	0.887***	0.947***
	[0.179]	[0.493]	[0.281]	[0.278]
Cash/TA	0.594***	1.059***	0.302**	0.411***
	[0.091]	[0.244]	[0.129]	[0.158]
TD/TA	-0.351***	-0.027	-0.423***	-0.619***
	[0.077]	[0.221]	[0.107]	[0.138]
Volatility	-1.102***	-1.179*	-1.338***	-1.145***
	[0.246]	[0.657]	[0.391]	[0.378]
Profitability	0.677***	0.127	0.542***	1.385***
	[0.101]	[0.264]	[0.126]	[0.273]
Size	0.071***	0.175***	0.078***	0.000
	[0.008]	[0.017]	[0.013]	[0.013]
RE/TA	0.001	-0.089**	0.001	-0.051
	[0.014]	[0.045]	[0.016]	[0.059]
SGR	0.304***	0.240**	0.341***	0.204*
	[0.046]	[0.096]	[0.061]	[0.109]
Constant	-1.431***	-3.512***	-1.596***	-0.584***
	[0.128]	[0.278]	[0.204]	[0.195]
Industry FE	Yes	Yes	Yes	Yes
Country FE	Yes	Yes	Yes	Yes
Observations	35,205	8143	12,504	14,558
Pseudo R-squared	0.029	0.045	0.022	0.013
Panel B:
Change in the probability of repurchase increase due to COVID	-5.4%	-1.4%	-6.3%	-5.5%

In all four sets of regressions, the coefficient estimates of control variables in general have signs as expected. Firms with higher risk and financial leverage are more (less) likely to decrease (increase) dividends or repurchases, while more profitable, larger firms with greater sales growth are less (more) likely to decrease (increase) dividends and repurchases. Accumulated retained earnings have significant impacts on dividend changes, but less on share repurchase.5

Overall, our regression analysis has confirmed the negative impact of COVID on corporate payout policy, i.e. firms are more likely to cut dividends or repurchases and less likely to increase dividends or repurchases after COVID hits. Moreover, we find evidence that cash holdings can help mitigate the negative impact of COVID in an international setting, complementing previous studies on corporate investments and performance for American firms (Tawiah and Keefe, 2022; Zheng, 2022).

Conclusion

In this paper, we examine the impact of the COVID-19 pandemic on corporate payout policy, for both dividends and share repurchases, of publicly listed firms in the G-7 countries. Our study shows that more (fewer) companies decrease (increase) their dividends or share repurchases after COVID started, but the scale of adjustments and the choice of payout channels vary across countries. Although corporate cash holdings helped mitigate the negative impacts of COVID on payout adjustments in North America and Japan, the impact was less significant for European firms.

One possible explanation for the widespread cut of dividends after COVID hit, and the lack of mitigating effect of corporate cash holdings in European countries, may lie with payout restrictions tied to COVID relief programs provided by governments. As detailed in Appendix 1, European countries have set strict restrictions on corporate payouts for companies receiving aid from some COVID relief programs, while such restrictions were less clear or did not exist for relief programs in the USA, Canada, and Japan. A more in-depth analysis would require researchers to identify firms that have received COVID aid, but this is beyond the scope of our paper. Regardless, we hope our study will motivate more research to explore cross-country differences in adjustments of corporate finance decisions in response to COVID, as well as the potential impacts of government aid on those differences.

CRediT authorship contribution statement

Christos Ntantamis: Methodology, Formal analysis, Writing – review & editing. Jun Zhou: Conceptualization, Methodology, Data curation, Software, Formal analysis, Project administration, Funding acquisition, Writing – original draft, Writing – review & editing.

Variable name	Description
Dividend Decrease (increase)	A firm-year observation is classified as Dividend Decrease (Increase) if the dividend amount decreases (increases) 1% or more from the previous year's dividend amount.
Repurchase Decrease (increase)	A firm-year observation is classified as Repurchase Decrease (Increase) if the share repurchase amount decreases (increases) 1% or more from the previous year's share repurchase amount.
Cash/TA	Cash and short-term investment divided by total assets
TD/TA	Total debt divided by total assets
Volatility	Stock return volatility measured as the standard deviation of monthly stock returns for a fiscal year
Profitability	Operating income divided by lagged total assets
Size	Natural logarithm of total assets in U.S. dollars
RE/TA	Retained earnings divided by total assets
SGR	Sales growth rate as measured by change in sales scaled by previous year's sales.