Shareholder payout in family firms

(1)

Shareholder payout in

family firms

A master thesis

for

MSc. International Financial Management at

University of Groningen

Name: Nikola Lazov Student Number: S3897141 Thesis supervisor: Adri de Ridder

Study programme: MSc International financial management

(2)

Abstract

(3)

1. Introduction

(6)

2 between family and non-family companies. Do family firms pay out more often than non-family firms? This paper aims to add to extant literature by drawing a comparison between the levels of payout in family and non-family firms.

For much of the 20th century, U.S. corporations, which can be regarded as the trend setters, have preferred to pay out dividends as opposed to other forms of shareholder compensation, largely due to the fact that it was the only known form of payout until around the end of the 1960’s (Grullon et al., 2002). The introduction of The Reform Act of 1986 is considered a turning point, which helped introduce and cement other forms of payout as a viable and sometimes preferable alternative to dividends such as stock repurchases. Dividends seem to hold a lesser role nowadays, with studies reporting a stagnation in their level increase, whereas a significant increase in share repurchase use has been recorded (Grullon et al., 2004). This gave rise to the substitution theory of dividends and repurchases and offered a way of explaining the dividend puzzle, introduced by Black (1976) and expanded upon by Fama and French (2001). Is there a difference in payout in family firms? By investigating the development of the payout policy of dividends and share repurchases in family firms, this paper aims to study the payout policy of family firms according to their life cycle and see if they conform to the norm.

(7)

3 to pay out, regardless of the payout vehicle chosen, a finding that is in line with the theory that family firms pay out more than non-family firms (Isakov et al., 2015).

The second question of the paper deals with the reliance of family firms on share repurchases and dividends as payout vehicles. By considering the possibility of the substitution effect in family firm payouts as well, this paper examines the payouts in family firms through the lens of the life-cycle theory and tests the effect of firm maturity on their payout strategy. Results show that the life-cycle theory of dividends does hold for family firms but repurchases do not, indicating that they should not be considered as substitutes. Further analysis shows that the family firm size does not affect dividend payouts and larger businesses are likely to pay out dividends just as often as smaller family owned firms. This finding indicates the importance of dividends to family firms and shows that this payout vehicle holds a significant place in the payout policy regardless of the popularity of share repurchases. Lastly, this study also extends a literature on family firms, payout policy, life cycle-theory in an international setting. The unique hand-collected sample adds to the research of family firms and helps to show that family firms have unique characteristics that set them aside from other companies.

(8)

4

2. Literature review

With increased interest in stock repurchases, the relationship between dividends and stock repurchases has become the subject of financial theory over the last 20 years. The relationship between the payout vehicles lies at the heart of payout policies and has been put to the test but a unanimous theory is yet to be accepted. Academic research has so far postulated two opposing points of view regarding the interaction between dividends and stock buybacks, as mentioned and expanded upon below.

At the ground level of theory, the Miller and Modigliani (1961) perfect world theory would suggest irrelevancy between dividends and share repurchases. When considering perfect and complete capital markets, dividends and share repurchases would be regarded as perfect substitutes, with no tax or market implications. The real world, however, introduces factors that may challenge this perception. While there are works that show a difference between share repurchases and dividends (see Jagannathan, Clifford and Weisbach (2000), Brav, Graham, Campbell and Michaely. (2005)), there are also those, who draw the conclusion of perfect substitution in accordance with the “perfect world” theory (see Grullon, Paye, Underwood and Weston (2011), Grullon and Michaely (2002) or Jiang, Kim , Lie and Yang (2013)). This debate has sparked research into the field of payout mix and produced many papers on the preference of firms regarding payout vehicles and frequency, such as Mensa et al. (2014), Grullon et al. (2008), Backwell, McWalter and Ritchken (2019) etc.

(9)

5

2.1. Development of payout policy and the substitution theory

According to Bierman (2002) and Baker (2002) dividends can be defined as “an appropriation of profits to shareholders after deducting tax and fixed interest obligations on debt capital”. Whereas share repurchases are characterized as a firm’s acquisition of a specified number of issued shares, usually announced through their share repurchase announcement in advance. There is an economic relationship to be observed between the two – stock repurchases affect the number of total shares outstanding and can thus lead to a change in the dividend per share ratio, improving company performance metrics. Share repurchases also have a signalling effect in showing a reported reduction in risk by the company, according to Grullon and Michaely (2004). This, in turn, could appeal to new investors, showing that the firm expects more stable cash flows in the future which could lead to higher payouts. Many academic papers, such as those by Grullon and Michaely (2002), Farre-Mensa, Michaely, and Schmalz (2014) and Skinner (2007), posit that share repurchases are gradually phasing out dividends and can be regarded as perfect substitutes in the eyes of the management and the market. Much of their research comes as a renewed response to the “dividend puzzle” question, introduced by Black (1976). This renewed interest was raised as a consequence of the paper by Fama and French (2001) and research is focused on the time period of 1978-2002: a period in the US corporate history that saw dividends plummet and the rise of share repurchases. From 1978 until 1996 the number of share repurchase programmes announced in the US increased six-fold to a value of more than US$ 110 billion. Dividends, on the other hand, had risen only from 67.6 billion to 141 billion US$. This meant that by 1999 share repurchases outgrew dividend payouts in value, prompting the question: are dividends being replaced by share repurchases?

(10)

6 show that with lower retained earnings, firms that were in a good position to pay out dividends, are not choosing to do so. A possible answer is provided by Grullon and Michaely (2002), who test the relationship between share repurchases and dividend payouts and conclude that the two vehicles for payouts can be regarded as substitutes. By looking at the Lintner model (Lintner, 1956) of dividend smoothing, they show a link between the increase in repurchase activity at the cost of potential dividend increases. Their research is, however, constrained by the assumptions of the Lintner model, which was derived at a time when share repurchases did not exist. Furthermore, research in 2011 by Grullon et al. (2011) looked at the relationship between net payouts and dividends. Results show that companies are currently distributing more net cash to shareholders than in the past and that share repurchases are the chief reason for a disproportionately large fraction of the variation in total payouts. These results are corroborated by von Eije and Megginson (2008) using a European sample. A possible reason for this is that dividends are stickier than share repurchases and can be attributed to the agency costs and signalling virtue of dividends (Chay and Suh, 2009). Brav, Graham, Harvey, and Michaely (2005) show that management of firms, who have steadily followed a payout policy of dividend distribution over a long period of time, are reluctant to reduce dividends and would prefer not to be paying dividends at all but feel compelled to do so by their firms’ dividend history.

(11)

7 Doukas, 2015) and outside shareholders in family firms are expected to accentuate that point, as they have little control over firm management. Furthermore, shareholders prefer the dividend vehicle as a payout method and the family ownership is a main driver for a higher likelihood of a dividend payout (Schmid, Ampeneberger, Kaserer and Achleitner, 2010). While the literature would then suggest that share repurchases are phasing out dividends and a substitution effect is being observed, the above-mentioned characteristics imply that share repurchases would play a lesser role in a family firm due to its increased reliance on dividends and their customary traits, their stickiness in particular.

2.2. Life-cycle theory of payout in family firms

A counter point to the substitution theory is the life cycle theory of dividends in family firms. The life-cycle theory of dividends, proposed by DeAngelo et al. (2006), states that the

probability that a firm pays out a dividend is dependent on its stage of maturity, measured by the mix of earned and contributed capital. They posit that firms, who find themselves in the earlier stages of their life cycle, proxied by a lower level of retained earnings to their total equity and with poor investment opportunities, refrain from paying out dividends. DeAngelo et al. (2006) support their theory using a US sample. Coulton and Rudock (2011) further test the life-cycle theory for dividends and share repurchases using an Australian sample and show that, in Australia, dividends remain the main payout vehicle and that repurchases have not yet approached the US level. They find evidence that retained earnings to total assets (RE/TA) significantly impacts the probability to pay out share repurchases as well and suggests that dividends and share repurchases should be interpreted as compliments, with the level of dividends being a determinant for the probability to initiate share buybacks. A possible interpretation of this theory can be found, when looking at Chay and Suh (2009), who

(12)

8 show that the certainty and uncertainty of cash flow determines payout policy and shows why some firms pay higher dividends than others, also contributing to the life-cycle theory. Their research is strengthened by Jagannathan et al. (2000), who extend the life-cycle theory of payout by introducing cyclicality. Their research finds share repurchases to be pro-cyclical and the result of temporary cash flows, while dividends are paid out form more permanent cash flows and extend the theory that repurchases are complementary to dividends and not substitutes. Building on this theoretical background, it is logical to assume that the life-cycle theory should hold for dividends and share repurchases in family firms as well. When the family firm reaches a certain stage of maturity, it should increase the level of its payout. The finding of Baker, Dam & De Ridder (2018) confirm the stickiness of dividends and proves that dividends in a European setting are still valued higher than share repurchases. They also show that higher retained earnings to contributed equity ratio increases the probability of a dividend payout.

2.3. Effect of firm size on payout policy

(13)

9 DeAngelo and Skinner (2007) that shows the preferences of a controlling shareholder as a first order impact on the dividend payout policy, it is reasonable to assume that larger family firms would be more likely to initiate payouts than smaller counterparts. In addition, larger firms are expected to have more agency problems to deal with and constraining the amount of cash at hand left for the CEOs to operate with is an important tool for dealing with agency problems for the shareholders (Farre-Mensa et al., 2014). This is usually achieved via dividend payouts. A connection between the firm size and the life-cycle theory of dividends is also evident from previous research and literature. Opler, Pinkowitz, Stulz and Williamson (1999) find that investment and growth opportunities dictate the levels of earnings retained at the firm in the form of cash, which would be available for payouts. According to their study, firms that are smaller in size tend to have higher growth opportunities and therefore retain less cash in comparison to their larger counterparts. This finding was corroborated by several studies, for example by the findings of Pinkowitz, Stulz and Williamson (2015), that higher levels of cash holdings are prompted by poor investment opportunities. Larger firms are, therefore, expected to have a higher level of retained earnings, which in turn leads to a higher stage of financial development in the life cycle of the firm.

3. Hypotheses

(14)

10

3.1. Propensity to pay of family and non-family firms

Based on the literature discussed above and the characteristics of dividend payouts and share repurchases the following hypotheses are derived. With firms exhibiting an increasing tendency to pay out earnings (Grullon et al., 2004) to shareholders via share repurchases and research showing a decreasing tendency to pay out via dividends (Grullon, 2011), the expectation for widely held shareholder firms is to have a higher probability of invoking a share repurchase programme instead of initiating a dividend payout. On the contrary, family firms exhibit unique characteristics that show them favouring dividend payouts instead of other payout vehicles. The stickiness of dividends is a main decisive factor, as the cyclicality of share repurchases makes them a less reliable vehicle for a steady income stream. As a large part of controlling shareholder families rely on the income from the firm as a form of personal wealth building (Isakov et al., 2015), especially in the early generations of the family starting and running the firm, the dividend stickiness is expected to prove a decisive factor. On this basis, the expectation is that the propensity of family firms to pay out via dividends should be higher than that of non-family firms. As an effect of that expected relationship, the contrary expectation is held for share repurchases. As family firms are expected to favour dividend-based distribution, they would engage less often in share repurchase programmes. Consequently, the propensity of share repurchase programmes for family firms is expected to be lower than that of non-family firms. Based on these unique firm characteristics, family firms payout policies are expected to differ from that on non-firms:

H1a: Family Firms tend to pay out more often in dividends than non-family firms

(15)

11

3.2. Life-cycle theory of payouts in family firms

An important determinant of the payout strategy of a firms is considered to be its life-cycle stage. The findings of DeAngelo et al. (2006) show a positive linear relationship between the proxy variable for life-cycle stage – retained earnings to total assets (RE/TA), and the

dividends payout of the firms, thus confirming the life-cycle theory of dividends. Other studies (Coulton and Rudock, 2011) have tested for a relationship between share repurchases and the life-cycle stage but found very limited support for that hypothesis. With stickiness stated as an important factor in deciding what payout strategy to choose in family-controlled firms and the positing of high importance of dividends in Europe (Baker et al., 2019), the expectance of life-cycle to apply in family firms for dividends is put forward. An aspect to consider is that, with the controlling family’s reliance on income from the owned company, the payouts are expected to be more frequent regardless of performance and stage of financial development. Building on that and considering the substitution theory of Grullon et al (2004), the assumption that both dividends, as well as share repurchases, should conform to the life-cycle theory is put forward. An alternative outcome for share repurchases is also considered, however, based on difference in cyclicality between share repurchases and dividends. Under the compliment theory (Jagannathan et al., 2000), it is possible to assume that share

repurchases do not conform to the life cycle stages of a family firm and are not used as substitutes for dividends. With the following hypothesis, the life-cycle theory for dividends and share buybacks in family firms is made:

H2a: Family firm’s life cycle increases the probability of dividends payout in family firms

(16)

12

3.3. Firm size effect on life-cycle theory of family firms

Expecting lower use of stock repurchases in family firms, this hypothesis aims to study the impact of the firm size on the payout strategy of family firms with different size. As there may be a big variance in the size of and composition of family firms, the moderating variable of market capitalization as a proxy for size (Redding, 1995) aims to help interpret the effect of life cycle of dividends and share repurchases in smaller family founded public firms compared to established market giants. The underlying assumption is, that as the firm grows, it starts to act more and more as a corporation. Firms in the same field often exhibit similar behaviour and competitive isomorphism (see Rodrigues and Craig 2007), which would imply that large family firms would seek to mimic the procedures of non-family firms. We therefore expect that firms with high market capitalization will have a positive effect on the relationship of share repurchases and firm maturity (RE/TA), as a mature firm will seek to pay out more of its earnings as a consequence of the fewer good investment opportunities for growth expected. A typical example of the variance in size expected in family firm operations size is the comparison of companies like Heineken and BMW to much smaller family enterprises like Summus Solutions. The positive difference in market capitalization for the former firms is expected to have a positive impact on the relationship of firm financial development stage and payout vehicles. Based on this, the following hypotheses are put forward:

H3a: Market capitalization positively moderates the relationship between firm maturity and dividend payout probability

(17)

13

4. Methodology

4.1. Institutional Setting

Literature and research on European countries regarding dividend and share repurchase payouts is not as extensive as the US one. This difference can be attributed to European regulation. Up until the year 2000, share repurchases were either prohibited or very restricted in many of the European Union countries (Sakinc, 2017). European regulation on share repurchases were introduced for the first time in the Second Council Directive 77/91/EEC of 13 December 1976 on 'coordination of safeguards' and it took more than two decades for them to gain momentum and become relevant in the corporate payout world (EEC, 1977). The EU introduced the Market Abuse Directive in 2003 to further regulate share buybacks and avoid their use as a tool, whose misuse could lead to maintaining the price of one or more financial instruments at an artificial level (de Matos, 2014). Firms from civil law countries, as is the case in continental European countries, have generally higher ownership concentration and can be characterized as having a lower information asymmetry between management and shareholders relative to common law countries (La Porta, 1998). This, however, does imply a lower investor protection level. Furthermore, the signalling effect of buybacks is expected to be of lower magnitude than in the US as a result (de Matos, 2014). The countries used in the sample are all part of the EU and the Eurozone and employ a code law system, so no difference in their payout strategy should be attributed to the law system.

(18)

14 tax regulation separately is required. Germany imposes the same rate for capital gains and for dividends at 25% (Einkommenssteuergesetz, 2019). France also has a unified policy, with taxes for both dividends and buybacks starting at 9.2% (Pricewaterhouse Coopers, 2019). An overview of Italy, released by Deloitte, also shows no difference in the tax rate (Deloitte, 2019). Netherlands uses a higher tax rate for capital gains than for dividend payouts, so the expectation here would be that Dutch investors would prefer the dividend as a payout vehicle (KPMG, 2019). Belgium, on the other hand, has a higher dividend tax than capital gains tax and this could affect the payout mix policy in favour of share repurchases (KPMG, 2019). Portugal offers the same rate for both payout vehicles at 28% ("Portugal - Income Tax", 2019). Finally, Spain offers the same rate for dividends and buybacks standing at 19% (KPMG, 2019). This shows that taxes do not play such a large role in the choice between the two payout vehicles and that motivation for the different payout choice can be sought elsewhere.

4.2. Classification of family firms

Following extant literature on the subject, a company is defined as widely held when there is no shareholder with a stake giving him 20% or more of the voting rights present (see Panunzi et al., 2006, Isakov and Weisskopf, 2015 etc.). Although 20% may come across as a low number, it is widely accepted to be the threshold for a controlling stake and enough for the shareholder to exercise influence over company decisions such as determining the payout policy. Some studies use a 50% threshold (Colot, 2015) but such a high-stake percentage limits the samples greatly and excludes firms who are effectively being run by their founding

(19)

15 as companies, where the founding family (or the combined stake of the founding families, in the case of multiple founders, as is the case with ALTRI SGPS S.A. with the founding families of Cannavaro and Mendoza) holds more than 20% of the company voting shares, regardless of the distribution between its different members. Furthermore, a requirement of a family member holding a position in management or on the board of directors or supervisory board is imposed, to ensure that the founding family takes active participation in the running of the company. Firms that meet the requirements above, are classified as family firms in the sample and a dummy variable is created to distinguish between them and widely held firms.

4.3. Sample

(20)

16 period of 20 years. Share repurchases were, if not prohibited, then difficult to implement in many European countries until the year 1998, so the time period should be set after that year (Sakinc, 2017). This international data will allow for the conduction of a comprehensive and robust test of the importance of stock repurchases and dividends for family firms in the European market and help enhance our understanding of the forces that affect corporate payout policy. The use of international data is a departure from prior studies of payout policies that generally focus on one market such as the U.S. The data was winsorized at the 1% confidence level to account for outliers. Furthermore, since some non-family firm observations exhibited very high negative values for retained earnings, the negative cap for negative retained earnings to assets was set at -0.5 and for retained earnings to total assets to -1. The stock price volatility data for both sets of firms was incomplete from the selected database and as such was omitted from the sample.

4.4. Variable definition

To test the effects of family firms on firm payouts, two different dependent variables are taken: 1. Dividends: defined as the total amount of dividends paid out on common stock

outstanding and normalized by the total value of assets. An alternative in extant literature is to use dividends scaled by earnings (see Grullon et al., 2011) but as earnings can also be negative, the value of total assets was chosen to avoid negative values for the dependent variables.

(21)

17 a proxy for stock repurchases is taken. Following Fama and French (2001), we employ the total dollar value change of Treasury stock as a proxy for net share repurchases. Treasury stock captures the cumulative effects of stock repurchases and reissues and it is not affected by new issues of stock (Fama and French, 2001). Changes in the treasury stock are multiplied by the share price to obtain the value of stock repurchases. Stock repurchase value is normalized by total assets.

Following extant literature, this paper follows DeAngelo et al. (2006) and Grullon et al. (2011) to control for the effects of firm characteristics. The following variables are included:

1. Market-to-book ratio (M/B): defined as the ratio of market value of equity to the book value of total assets

2. Return on assets (ROA): defined as the operating income before depreciation scaled by total assets

3. Sales growth rate (SGR): defined as the change in total sales from one accounting period to the next

4. Retained earnings to total assets (RE/TA): This variable is equal to retained earnings scaled by the book value of assets

5. Retained earnings to total equity (RE/TE): This variable is defined as the value of retained earnings scaled by the value of total equity

6. Total equity to total assets (TE/TA): This variable is defined as the total value of equity scaled by the total value of assets.

(22)

18 8. Firm size (MCap): firm size is defined as the market capitalization of the firm scaled by

the value of total assets

9. Cash (Cash/Assets): The cash holdings of the firm are defined as the amount of cash reserves scaled by total assets

10. Leverage (Lev): To account for the capital structure of the firm and how it impacts payout strategy, we use leverage, defined as the debt to equity ratio.

4.5. Regression model

4.5.1 Propensity model

Hypothesis 1 is tested using the propensity score matching (PSM) method. Following Grullon et al (2011), the hypothesis is tested by measuring the difference in the propensity of dividend payouts and share repurchases in family and non-family firms. To do so requires the firms in the control group – the non-family firms, and the target group – family firms, to be matched according to significant characteristics (covariates). This can be done after the propensity to pay out via the selected vehicle has been calculated, using equation (1). Finally, the difference in their propensity score will yield the result of the effect on the family firms as the treatment group. A logit model is used for this calculation. This is achieved in three steps:

1) Pr(Yj,t=1) = f(β’Xj,t), (1) where Y is a variable equal to 1 if the firm pays, f() represents the logistic function and X is a vector of covariates.

2) Match the participants using the estimated scores

(23)

19

4.5.2 Life-cycle model

To test hypothesis 2, following extant literature (see DeAngelo et al., 2006, Coulton et al., 2011, Baker et al., 2018), the Fama and MacBeth (1973) statistical methodology is employed to test whether the probability of a family firm paying dividends or share repurchases exhibits systematic dependence on the stage of the firm’s financial life cycle. A logistic regression is used to test the relationship between the payout vehicles and life-cycle theory. Following Baker (2004), retained earnings to total assets is taken as a proxy for maturity. This yields the following equation:

Pr(Dividendsj,t = 1) = β0 + β1*RE/TAj,t + β2*controlsj,t + εj,t , (2) where Pr(Dividendsj,t = 1) is the probability of paying out dividends. RE/TAj,t is the independent variable. The control variables are denoted by controlsj,t and include profitability(ROA), asset growth, leverage, cash, sales growth and investment opportunities (M/B ratio). εj,t is the error term. The same equation is used to calculate the possibility of a relationship between the life cycle of a family firm and the probability to initiate share repurchases with the only difference being the dependent variable. It is denoted as Pr(Stock repurchasesj,t = 1) and represents the probability of initiating share repurchases:

Pr(Stock repurchasesj,t = 1) = RE/TAj,t + controlsj,t + εj,t (3)

4.5.3 Maturity model

(24)

20 purpose, the dummy variable HighMCap is created. It takes on the value of 1 if the firm is in the 67th percentile or higher and 0 otherwise.

To test the final hypothesis of the difference of payouts in family firms according to the stage of development they are in, a moderating variable is defined as:

HighMCap*RE/TA – This variable is equal to the return on assets multiplied by retained earnings scaled by the book value of assets and used as a proxy for the performance of family firms according to their stage of maturity in the life-cycle of the firm.

This yields the following equation:

Pr(Yj,t = 1) = β0 + β1*RE/TAj,t + β2*RE/TAj,t*HighMCapj,t + β3*HighMCapj,t + β4*controlsj,t+ εj,t , (4)

Where Yj,t is the event of a dividend payout, RE/TAj,t is the proxy for the life-cycle stage of the firm and the HighMCap is the proxy for firm size, which takes on the value of 1 when the firm is above the 67th percentile in the market capitalization parameter of the sample and 0 otherwise. The equation remains the same for share repurchases with the note that in this case Pr(Yj,t = 1) represents the conduction of a share repurchase programme.

4.6. Descriptive statistics

(25)

21 Table 1 shows summary statistics for different variables, split by firm ownership. The average amount of dividends paid for non-family firms is 0.01 with a 0.25 standard deviation. In contrast, family firms show a higher mean for dividend payout, as was expected – with a 0.02 mean and a much smaller deviation at 0.04. This implies that dividends are stickier in family firms and is in line with the finding of Isakov and Weisskopf (2015) that family firms tend to pay out more. Share repurchase for family firms also have a nominally higher value at 0.002 in comparison to 0.001. This surprising finding would suggest that family firms pay out a similar if not higher total amount of dividends and share repurchases on average. While this is surprising for buybacks, it does support the idea that payouts in founding family firms are made more often and are higher in relative numbers, considering that both payout vehicles demonstrate higher means in comparison. Non-family firms are expected to be bigger and have more assets and this is corroborated by the variables retained earnings to assets (RE/TA) and retained earnings to equity (RE/TE). Those two variables have a much higher mean and a lower standard deviation for family firms, which indicates a smaller assets base and more stable operations. A likely explanation for this is that family firms have lower agency cost and information asymmetry (Isakov and Weisskopf, 2015), which may be the reason for having a higher mean of RE/TA. A further indicator that family firms may be smaller than their counterparts in general can be seen by looking at the market to book ratio and sales growth. The higher market to book mean of 1.84 (compared to 1.75) and the higher sales growth ratio of 1.06 (compared to 1.04) indicate that family firms still have much higher growth

(26)

22

Non-family firms Family firms

Quantiles Quantiles _Wilcoxon

rank-sum Variable N Mean S.D. Min .25 Mdn .75 Max n Mean S.D. Min _{.25 Mdn .75 Max}

Retained earnings/assets 23315 0.05 0.27 -0.50 -0.02 0.05 0.20 0.71 4215 0.14 0.25 -0.50 _{0.02 0.11 0.29 0.71 -19,229***} Retained earnings/equity 23315 0.30 0.94 -1.00 0.00 0.20 0.65 6.52 4215 0.43 0.81 -1.00 _{0.08 0.36 0.81 6.52 -16,482***} Total equity/assets 23315 0.39 0.23 0.00 0.22 0.38 0.54 0.96 4215 0.40 0.20 0.00 _{0.25 0.39 0.54 0.96 -3,891***} Market to book ratio _{23315 1.75 1.17 0.00} _0.96 _{1.37 2.47 4.00} _{4215 1.84 1.15 0.00} _{1.00 1.48 2.59 4.00 -5,515***} Return on assets _{23315 0.00 0.16 -0.87} _0.00 _{0.03 0.07 0.34} _{4215 0.04 0.11 -0.87} _{0.02 0.05 0.08 0.34 -21,202***} Sales growth _{23315 1.04 0.59 0.00} _0.92 _{1.03 1.15 4.62} _{4215 1.06 0.49 0.00} _{0.97 1.06 1.15 4.62 -5,921***} Cash 23315 0.09 0.13 0.00 0.00 0.04 0.11 0.69 4215 0.09 0.10 0.00 _{0.02 0.06 0.12 0.69 -7,946***} Leverage _{23315 0.11 0.24 -0.42} _0.01 _{0.05 0.12 1.53} _{4215 0.10 0.18 -0.42} _{0.02 0.05 0.12 1.53 -3.937***} Size _{23315 0.99 1.22 0.02} _0.32 _{0.66 1.05 8.10} _{4215 1.02 1.19 0.02} _{0.37 0.69 1.10 8.10 -4.298***} Dividends _{23315 0.01 0.25 0.00} _0.00 _{0.00 0.02 0.17} _{4215 0.02 0.04 0.00} _{0.00 0.01 0.02 0.17 -21,137***} Share repurchases 23315 0.00 0.01 0.00 0.00 0.00 0.00 0.03 4215 0.00 0.00 0.00 0.00 0.00 0.00 0.03 -8,646*** Table 1 Univariate analysis

(27)

23 Dividends, on the contrary, have a more even distribution but the largest dividend payers are again, expectedly, at the top end, with a value of 0.0401 at the 90th percentile. This statistic serves to show that family firms continue to rely on dividends as a main payout vehicle and may start to act like non-family firms only when they reach a certain level of financial development and maturity.

Graph 1 shows a bar plot representation of the family firms that pay out and their payout vehicles by country. The overview confirms that, in all countries selected, a higher percentage of family firms pay out via dividends rather than by share repurchases. Germany and Portugal firms show the highest taste for dividends, whereas a higher percentage of French and Spanish family firms buy back shares in comparison to others. Despite having more economically beneficial conditions for share repurchases with regard to taxation of payout, more than 70% of Belgian family firms pay out dividends and only 25% of firms choose to use share repurchases as well.

Graph 1 Family firm payout by country

The graph shows the percentage of firms that pay out dividends and buy back shares according to their country of origin

(28)

24 The decision to pay out more in dividends than in share repurchases can be attributed to the preferences of the founding family shareholders and their special taste for dividends (Schmid et al., 2010). Table 3 shows the summary statistics of dividends and share repurchases in family firms by country. It corroborates the information in graph 1 that dividends are the more common payout vehicle in family firms. It depicts the strong preference of Germany and Portugal towards dividends with a mean of 0.142 and 0.147 respectively. Both countries have a very low mean for share repurchases below the 0.1% value, which is markedly different from the rest of the sample, where share repurchases have a mean of 0.1%. The low number of the Portuguese sample of 13 firms shows that share repurchases within the sample from that country are almost non-existent. Looking at table 5, it can be seen that a possible reason for that is that the size of Portuguese family firms is much smaller than the rest of the sample with a mean of 0.492. This could explain the low interest Portuguese firms show in share repurchase programmes. A higher number of bigger companies can be found in Germany and the Netherlands. Germany is one of the strongly represented countries, together with France, with over 70 companies represented. Considering a strong cultural attachment in these countries towards family businesses (La Porta, 1998), it comes as no surprise that the

dividend payouts in both have a much higher mean than the share repurchases, both having a dividend mean higher than 10%. These statistics are in line with the finding of Baker et al. (2018) that European firms still show a strong preference for dividends.

(29)

25 stemming from the founding family. Germany, shows the lowest family to board members ratio, indicating a lower investment in the company at first, but this can be attributed to the larger size of the companies (see table 5) and the need for higher board membership.

Graph 2 Percentage of family members on the board per country in family firms

(30)

26

5. Empirical findings

5.1. Findings on propensity to pay out dividends and share buybacks

The first part of this section deals with the comparison in payout behaviour of comparable public family run and widely held firms. The firms were matched according to several characteristics and the probability of each firm to use either of the two payout vehicles. The difference in the probability of the matched firm couples yields the results to answer

hypotheses 1a and 1b. Table 6 summarizes the findings on propensity to pay out for family firms in comparison to non-family firms from equation (1). It shows the values of the coefficients of the covariates and the difference between the treatment and the control group in the propensity to pay out (Difference). Several specifications of a propensity score

(31)

27

Table 6 Regression results for hypothesis 1

Columns (1) and (4) show results from propensity score matching using a logit regression and including only profitability and the basic life-cycle proxy as covariates. Columns (2) and (5) show results from propensity score matching using a logit regression including the covariates for growth opportunities (SGR, M/B ratio), cash(Cash/Assets), capital structure (leverage) and the firm size (market capitalization/total assets). Columns (3) and (6) show results using a logit regression model which includes all covariates from the life-cycle theory of dividends. The dependent variable Dividend is a dummy variable that takes on the value of 1 in the event of a dividend payout and 0 otherwise. The dependent variable share repurchases is a dummy variable that take on the number 1 in case of an initiation of buybacks and 0 otherwise. *, ** and *** indicate significance levels at the 10%, 5% and 1% respectively.

and adding leverage and operating cash, following Fama and French (2001), in column 2, shows that M/B ratio, operating cash(cash/assets) and leverage also are statistically significant determinants of the likelihood of dividend payout. Sales growth does not hold statistical significance in any of the specifications. Column (3) shows the full model including all covariates for the life-cycle model of dividends (RE/TA, RE/TE and TE/TA), following DeAngelo et al. (2006). The results show that family firms have a 9.2% higher propensity to pay out dividends than non-family firms. Sales growth, operating cash and firm size appear to not have statistically significantly impacted the calculation of the propensity. The strongest impact on the difference in calculated propensity is had by profitability (ROA) and the ratio of retained earnings to total assets (RE/TA) with a higher increase in the probability to pay out of 1.615 and 0.844 respectively. This effect could be attributed to family firms having lower agency cost and information asymmetry (Isakov and Weisskopf, 2015) and a lower propensity

Ffirm Dividends Repurchases

(1) (2) (3) (4) (5) (6) RE/TA 0.856*** 0.844*** 0.844*** 0.856*** .844*** 0.843*** ROA 1.683*** 1.596*** 1.615*** 1.683*** 1.596*** 1.615*** M/B ratio 0.046*** 0.047*** .046*** 0.047*** SGR 0.019 0.025 0.019 0.025 Cash/Assets -0.307** -0.230 -0.307** -0.230 Leverage -0.254*** -0.300*** -0.254*** -0.300*** Size -0.018 -0.005 -0.018 -0.005*** RE/TE 0.077*** .0774*** TE/TA -0.301*** -0.301*** constant -1.836*** -1.864*** -1.797*** -1.836*** -1.864*** -1.797*** Difference 0.099*** 0.084*** 0.092*** 0.052*** 0.052*** 0.052*** Common support

(32)

28 to invest in negative net present value projects. A family run firm is expected to have more risk-averse CEOs and may seek to maximize utility rather than wealth, as proposed by Caliskan et al. (2015) regarding risk-averse CEOs. Leverage is statistically significant and negative of sign at -0.3, showing the negative impact debt has on the probability of payouts. Furthermore, the lower mean of leverage for family firms would also indicate a lower financial constraint for family firms, which would enable them to pay out more often. As a summary of the effect on dividends, higher profitability and investment opportunities, a higher personal stake and lower leverage have high significantly explanatory power in the determining of the likelihood of dividend payout. Furthermore, dividends payouts are 9.2% more likely to occur in a family firm than a widely held public firm.

Columns (4), (5) and (6) in table 6 exhibit the results for share repurchases. All 3 specifications show significant results and report a higher likelihood of a share buyback programme occurring in family firms than in non-family firms. As with dividends, the strongest impact on the calculation of the propensity difference can be found in the

(33)

29 stems from the difference in profitability and retained earnings in family firms. These results contradict and disprove hypothesis 1b and highlight a higher propensity of a family firm to initiate buybacks compared to a matched counterpart. This result, though unexpected, is in line with literature on family firms (Schmid et al., 2009) that family firms exhibit a higher propensity for both dividends and total payouts. It also lends limited support to the

complement theory, as it shows that a higher propensity to utilize one form of payout does not lead to a lower propensity to utilize the other. This is in contradiction of the substitution theory of the interchangeability of dividends and share repurchases, on the assumption of which hypothesis 1 was derived. Furthermore, share repurchases are a procyclical form of payout and, as such, are expected to be initiated in favourable conditions. On one hand, they might be used in times of excess temporary cash flows to distribute fluctuating earnings (Jagannathan et al., 2000). Alternatively, in times of lower share price, managers might initiate share buyback programmes as a way to boost performance metrics and capitalize on the lower price (Jagannathan et al., 2000). Furthermore, family run management might use such times to initiate buybacks as a way to regain larger control of its firm, forcing periphery shareholders out of the firm (Isakov and Weisskopf, 2015).

5.2. Life cycle theory of dividends and share repurchases

To see whether the likelihood of a payout is determined by similar factors as in non-family firms, the life cycle theory - put forward by DeAngelo et al. (2006) - is tested. The theoretical impact of the firm’s stage of financial development is tested for both dividends and

(34)

30 toward contributed capital mix, following DeAngelo etl al. (2006). A similar structure has been used for share repurchases in columns (4), (5) and (6).

Table 7 Regression results for the life cycle theory of dividends and share repurchases in family firms

Columns (1), (2) and (3) use a Fama and MacBeth regression with Dividend Payout as the dependent variable. The dependent variable Dividend is a dummy variable that takes on the value of 1 in the event of a dividend payout and 0 otherwise. Columns (4), (5) and (6) use a Fama and MacBeth regression with Share repurchases as the dependent variable. The dependent variable share repurchases is a dummy variable that take on the number 1 in case of an initiation of buybacks and 0 otherwise. The independent variables RE/TA and RE/TE and the control variables are explained in section 3. *, ** and *** indicate significance levels at the 10%, 5% and 1% respectively.

Fama-MacB Dividends Repurchases

(1) (2) (3) (4) (5) (6) ROA 0.008*** (0.001) 0.012*** (0.001) 0.012*** (0.001) 0.000 (0.001) 0.000 (0.001) 0.000 (0.001) RE/TA 0.223*** (0.050) 0.242*** (0.051) 0.004*** (0.095) 0.068 (0.022) 0.075*** (0.023) 0.168 (0.108) TE/TA -0.336*** (0.051) -0.208*** (0.070) -0.153*** (0.052) -0.184** (0.081) RE/TE 0.120** (0.050) -0.031 (0.042) Asset growth 0.131*** (0.039) 0.116*** (0.039) 0.124*** (0.039) 0.060 (0.038) 0.050 (0.003) 0.050 (0.003) Leverage -0.181 (0.066) -0.201** (0.088) -0.004** (0.091) 0.181** (0.084) 0.083 (0.082) 0.071 (0.109) Cash/Assets 0.246*** (0.064) 0.283*** (0.068) 0.254*** (0.054) 0.258** (0.109) 0.277** (0.110) 0.261** (0.111) Sales growth -0.054 (0.037) -0.064* (0.034) -0.070* (0.034) -0.010* (0.037) -0.011 (0.037) -0.012 (0.037) M/B ratio 0.008** (0.003) 0.007** (0.003) 0.010*** (0.003) 0.001 (0.005) 0.001 (0.005) 0.002 (0.005) Constant 20.222*** (1.883) 18.764*** (1.800) 18.535*** (1.620) 0.680 (2.964) 0.190 (2.954) - 0.201 (2.844) R2 0.256 0.277 0.301 0.062 0.072 0.078

(35)

31 negative impact on dividend payout probability. TE/TA is introduced as a control for the effect of the leverage and it’s negative sign corroborates the findings of DeAngelo et al. (2006) that earned/contributed capital mix and not the total equity relative to other sources of capital is a key determinant of the dividend payout decision. This may serve as an indicator that the firm’s earned/contributed capital mix plays a vital role with the sign on RE/TE being positive in column (3). This could suggest that contributed capital does not lead to payouts, while a preference for earned capital to be paid out is clearly shown. As expected from

literature, a higher M/B ratio marginally increases the probability of a dividend payout (0.01), while having good investment opportunities and recording growth, as proxied by sales

growth, reduces the dividend payout likelihood, as cash is reinvested in the firm. The

coefficient for sales growth is significant at the 10% level and negative at -0.064, showing the strong impact of growth opportunities on the decision to pay out. Leverage also has a negative impact, an indicator that highly leveraged family firm CEOs would be encouraged to redirect firm cash towards improving the financial situation of a firm, rather than decreasing it’s assets by extracting funds from the firm. Looking at the complete model for dividends in column (3) shows the strong support for earned/contributed capital mix (RE/TE) as a determinant of dividend payouts at 0.120 odds increase per RE/TE unit increase. It can be inferred from the high coefficient and the now lowered coefficient of RE/TA (0.004) that the return on equity (ROE) rather than the return on assets (ROA) is of greater importance in family firms as a measure of profitability when considering dividends. These results provide empirical support for hypothesis 2a that the family’s firm life cycle is positively related to the probability to pay out dividends.

(36)

32 as compliments to dividends (Coulton et al., 2011). The findings in column (5) alone show a statistically significant impact of RE/TA to the probability of initiating a share repurchase programme. A look at table 2 shows that share repurchases occur only in the 90th percentile in family firms. Coupled with the result of table 6, it serves as evidence that financial maturity of the family public company leads to an increased probability of it employing share buybacks as a payout vehicle. A possible explanation behind this finding could include that once the firm reaches a mature life cycle stage and operates with more stability, the controlling family stakeholder may look to initiate buybacks to ascertain control and remain as a majority shareholder. When RE/TE is added to the equation in column (6) though, as seen in column (4) as well, the life cycle variables become statistically insignificant. Column (6) shows that the statistical significance of RE/TA in the previous specification of the equation can be due to an omitted variable bias and not as a support for the life cycle theory of share repurchases. This result contradicts hypothesis 2b that share repurchases are positively affected by the life cycle of the firm and contradicts the substitution theory of Grullon et al. (2002) in family firms. It adds to the understanding that share repurchases are rather cyclical and sporadic and are not determined by the stage of development of the firm. It does provide limited support to the theory, however, that dividends and share repurchases can be viewed as complementary payout vehicles (see Jagannathan et al., 1999).

5.3. Maturity of a family firm

(37)

33 (MCap*RE/TA) of market capitalization scaled by total assets (MCap) and the life cycle proxy RE/TA is statistically insignificant for both dividends and share repurchases. Focusing on column 1, the life cycle proxy variables RE/TA and RE/TE are both statistically significant at the 1% level, which again supports the hypothesis of life cycle theory of dividends in family firms. An increase in one unit of the earned/total capital mix (RE/TA) of small and medium sized family firms increases the odds of a dividend payout by 2.468. The interaction variable, representing the impact of the earned/total capital mix (RE/TA) of large family firms is statistically insignificant, showing that the large family firms do not behave differently than

Table 8 Regression results for the moderating effect of firm size on the life-cycle

Column (1) uses a logit regression with Dividends as the dependent variable. The dependent variable Dividend is a dummy variable that takes on the value of 1 in the event of a dividend payout and 0 otherwise. Column (2) uses a logit regression with Share repurchases as the dependent variable. The dependent variable Share repurchases is a dummy variable that takes on the value of 1 in the event of a share buyback and 0 otherwise. The independent variable RE/TA represents the life cycle stage of the firm and the interaction variable MCap*RE/TA represents the interaction between firm size and life cycle stage. *, ** and *** indicate significance levels at the 10%, 5% and 1% respectively.

logit Dividends Repurchases

(38)

34 their smaller counterparts. This finding is contradictory to the hypothesis 3a that family firms start to behave more like non-family firms as they mature and grow. Grullon et al. (2002) remark that the dividend payouts are increasingly focused in a smaller sample of big

companies that pay large dividend payouts. This, however, does not appear to be the case for family firms. As shown by the univariate analysis and by table 8, family firms pay out dividends regardless of their size. The findings are consistent with DeAngelo et al. (2006), who posit that an increase in the firm’s stage of financial maturity leads to an increase in the likelihood of a dividend payout. The firm size, however, does not seem to play a major role. Small-scale family firms, who are in a mature stage of the life cycle of their financial

development, would exhibit the same probability of a dividend payout as a large-scale family firm in the same stage. As expected, and shown in table 7, share repurchases do not conform to the life cycle theory and the independent (RE/TA) and interaction variable are both statistically insignificant in table 8, indicating that, regardless of size, family firms do not consider share repurchases as a substitute of dividend payouts. This reneges hypothesis 3b. Available cash is statistically significant, however, as it is in table 7, column (6). With a coefficient of 1.891, a possible reason for this result is that the increase of available cash increases the probability of the controlling shareholder considering a share repurchase programme as a way to consolidate their control (Bertrand, Mehta and Mullainathan 2002).

6. Conclusions

The objective of this paper is to examine the probability of shareholder payouts in a family firm context. By matching the sample of family firms with non-family firms, evidence was obtained that public firms - run by their founding family - exhibit a higher propensity for payout, regardless of the payout vehicle chosen. This confirms the findings of existing

(39)

35 firm dividend payouts and share repurchases, results suggest that the theory does not apply for both payout vehicles, with the financial stage of development proxy of retained earnings to assets being significant and positive for dividends only. These results contradict the

substitution hypothesis, put forward by Fama and French (2001) and elaborated by Grullon et al. (2002), by showing that share repurchases do not change with the life cycle stage of the company. A possible explanation of this can be found in the alternative theory, which suggests that they are used rather as compliments to dividend payouts (Coulton et al., 2011) and are implemented in family firms only once the company reaches a certain size and maturity stage as a control mechanism.

The results of hypotheses 2 and 3 show that family firms do not differ in their payout that much from other firms. As with widely held firms, dividends in family firms are affected by the life cycle theory and share repurchases are not considered as substitutes in that regard. Rather interestingly and contrary to the findings of Grullon et al. (2002), the probability of dividend payouts does not increase with the size of the firm. While the dividend payers in public firms may be clustered in the hands of a few large companies (Grullon et al. ,2002), the findings of hypothesis 3 show that the dividends payers among family companies are more spread out.

The stickiness of dividends and their signalling power can be interpreted as two important factors that contribute to the higher probability of payout in family firms. The lower agency costs, as seen by the higher mean of retained earnings in family firms as compared to widely held companies, in family firms also mean that less cash is invested in projects with low returns and thus more retained earnings remain available for distribution amongst the

(40)

36 personal income (Isakov et al., 2015). Given the scarce literature available on payouts in family firms, a possible topic for future research would be to examine the payout propensity of family firms compared to widely held firms in the different generations of the controlling families. Despite their cyclicality, family firms show a higher propensity to initiate share repurchases than widely held firms. This result can serve as the basis of future research examining the reasons behind it. A possible explanation is that share repurchases are used by the controlling family as a stake control mechanism in order to preserve control. Another explanation could be found in the large variance in the sample of family firms. Family companies do not use repurchases at all in some of the countries, such as Portugal, in the sample but do use them in others, such as Spain and France, for example.

This paper contributes to the existing body of literature on family firms by examining the payout policies and probability of payout. One of the main contributions of the paper is the use of a unique hand-collected international sample, which allows to account for a country bias and show a more complete picture of family firm behaviour. The paper adds to previous research (Schmid et al., 2010) in proving that family firms exhibit a higher probability of both dividend payout and share repurchase initiation. It also adds to previous research (DeAngelo et al., 2006) by examining the life cycle theory of both dividends and share repurchases in the context of a family firm. The lack of data on share repurchases in both databases Eikon and WRDS forced the use of a proxy (change in treasury stocks) for share repurchases and this needs to be considered as a limitation of the conducted research.

(41)

37

References

Andres, C. (2011). Family ownership, financing constraints and investment decisions. Applied

Financial Economics, 21(22), 1641-1659.

Andres, C., & Hofbaur, U. (2019). In for a Bumpy Ride? Cash Flow Risk and Dividend Payouts. SSRN Electronic Journal. doi: 10.2139/ssrn.3461574

Backwell, A., McWalter, T., & Ritchken, P. (2019). On Buybacks, Dilutions, Dividends, and the Pricing of Stock-Based Claims. SSRN Electronic Journal.

Baker, H. (2002). Increasing Shareholder Value: Distribution Policy, A Corporate Challenge.

Journal Of Financial Research, 25(2), 302-303.

Baker, H., Powell, G., & Veit, E. (2002). Revisiting the dividend puzzle. Review Of Financial

Economics, 11(4), 241-261.

Bertrand, M., Mehta, P., & Mullainathan, S. (2002). Ferreting out Tunneling: An Application to Indian Business Groups. The Quarterly Journal Of Economics, 117(1), 121-148.

Bhattacharya, S. (1979). Imperfect Information, Dividend Policy, and "The Bird in the Hand" Fallacy. The Bell Journal Of Economics, 10(1), 259.

Black, F. (1976). The Dividend Puzzle. The Journal Of Portfolio Management, 2(2), 5-8.

Bodnaruk, A., & Östberg, P. (2013). The Shareholder Base and Payout Policy. Journal Of

Financial And Quantitative Analysis, 48(3), 729-760.

Brav, A., Graham, J., Harvey, C., & Michaely, R. (2004). Payout Policy in the 21st Century.

Journal Of Financial Economics, 77(3), 483–527.

Caliskan, D., & Doukas, J. (2015). CEO risk preferences and dividend policy decisions.

Journal Of Corporate Finance, 35, 18-42.

Chay, J., & Suh, J. (2009). Payout policy and cash-flow uncertainty. Journal Of Financial

(42)

38 Colot, O. (2015). Family firms performance in Belgium. International Advances In Economic

Research, 21(2), 245-246.

Coulton, J., & Ruddock, C. (2011). Corporate payout policy in Australia and a test of the life-cycle theory. Accounting & Finance, 51(2), 381-407.

De Matos, J. (2020). Share Repurchases in Europe: A View over Managerial Hubris, EU Regulations and Country Legal Origins. Retrieved 8 December 2019, from

http://arno.uvt.nl/show.cgi?fid=134651

DeAngelo, H., & DeAngelo, L. (2006). Capital Structure, Payout Policy, and Financial Flexibility. SSRN Electronic Journal.

DeAngelo, H., DeAngelo, L., & Skinner, D. (2007). Corporate Payout Policy. Foundations

And Trends® In Finance, 3(2–3), 95-287.

DeAngelo, H., DeAngelo, L., & Stulz, R. (2006). Dividend policy and the earned/contributed capital mix: a test of the life-cycle theory☆. Journal Of Financial Economics, 81(2), 227-254.

EEC. (1977). Second Council Directive of 13 December 1976. Official Journal of European Communities.

EStG - nichtamtliches Inhaltsverzeichnis. (2019). Retrieved 9 December 2019, from https://www.gesetze-im-internet.de/estg/

European Union. (2017). Monthly Bulletin May 2017 (p. 2). European Union. Retrieved from https://www.ecb.europa.eu/pub/pdf/other/pp103-111_mb200705en.pdf

Fama, E., & French, K. (2001). Disappearing dividends: changing firm characteristics or lower propensity to pay?. Journal Of Financial Economics, 60(1), 3-43.

Fama, E., & MacBeth, J. (1973). Risk, Return, and Equilibrium: Empirical Tests. Journal Of

Political Economy, 81(3), 607-636.

Farre-Mensa, J., Michaely, R., & Schmalz, M. (2014). Payout Policy. Annual Review Of

(43)

39 Grullon, G., & Michaely, R. (2000). The Information Content of Share Repurchase Programs.

The Journal Of Finance, 59(2), 651-80.

Grullon, G., & Michaely, R. (2002). Dividends, Share Repurchases, and the Substitution Hypothesis. The Journal Of Finance, 57(4), 1649-1684.

Grullon, G., Paye, B., Underwood, S., & Weston, J. (2012). Has the Propensity to Pay Out Declined?. Journal Of Financial And Quantitative Analysis, 46(1), 1-24.

Huang, M., Li, P., Meschke, F., & Guthrie, J. (2015). Family firms, employee satisfaction, and corporate performance. Journal Of Corporate Finance, 34, 108-127.

Isakov, D., & Weisskopf, J. (2015). Pay-out policies in founding family firms. Journal Of

Corporate Finance, 33, 330-344.

Italy Highlights 2019. (2019). Retrieved 8 December 2019, from

https://www2.deloitte.com/content/dam/Deloitte/global/Documents/Tax/dttl-tax-italyhighlights-2019.pdf?nc=1

Jiang, Z., Kim, K., Lie, E., & Yang, S. (2013). Share repurchases, catering, and dividend substitution. Journal Of Corporate Finance, 21, 36-50.

John, K., & Williams, J. (1985). Dividends, Dilution, and Taxes: A Signalling Equilibrium.

The Journal Of Finance, 40(4), 1053-1070.

KPMG. (2019). Income tax Belgium. KPMG. Retrieved from

https://home.kpmg/xx/en/home/insights/2011/12/belgium-income-tax.html

KPMG. (2019). Income Tax Netherlands. KPMG. Retrieved from

https://home.kpmg/xx/en/home/insights/2011/12/Netherlands-income-tax.html

KPMG. (2019). Income tax Spain. KPMG. Retrieved from

https://home.kpmg/xx/en/home/insights/2011/12/spain-income-tax.html

La Porta, R., Lopez‐de‐Silanes, F., Shleifer, A., & Vishny, R. (1998). Law and Finance.

(44)

40 Lasfer, M., & Andriosopoulos, D. (2011). The Market Valuation of Share Repurchases in Europe. A Cross Country Analysis. SSRN Electronic Journal.

Miller, M., & Modigliani, F. (1961). Dividend Policy, Growth, and the Valuation of Shares.

The Journal Of Business, 34(4), 411.

Miller, M., & Rock, K. (1985). Dividend Policy under Asymmetric Information. The Journal

Of Finance, 40(4), 1031-1051.

Opler, T., Pinkowitz, L., & Williamson, R. (1999). The determinants and implications of corporate cash holdings. Journal Of Financial Economics, 52(1), 3-46.

Pinkowitz, L., Stulz, R., & Williamson, R. (2015). Do U.S. Firms Hold More Cash than Foreign Firms Do?. Review Of Financial Studies, 29(2), 309-348.

Portugal - Income Tax. (2019). Retrieved 9 December 2019, from

https://home.kpmg/xx/en/home/insights/2011/12/portugal-income-tax.html

Pricewaterhouse Coopers. (2019). France - Taxes on personal income. Pricewaterhouse Coopers. Retrieved from http://taxsummaries.pwc.com/ID/France-Individual-Taxes-on-personal-income

Redding, L. (1997). Firm Size and Dividend Payouts. Journal Of Financial Intermediation,

6(3), 224-248.

Rodrigues, L., & Craig, R. (2007). Assessing international accounting harmonization using Hegelian dialectic, isomorphism and Foucault. Critical Perspectives On Accounting, 18(6), 739-757.

(45)

41 Schmid, T., Ampenberger, M., Kaserer, C., & Achleitner, A. (2010). Controlling Shareholders and Payout Policy: Do Founding Families Have a Special 'Taste for Dividends'?. SSRN

Electronic Journal. doi: 10.2139/ssrn.1553650

Skinner, D. (2008). The evolving relation between earnings, dividends, and stock repurchases.

Journal Of Financial Economics, 87(3), 582-609.

Stephens, C., Jagannathan, M., & Weisbach, M. (1999). Financial Flexibility and The Choice Between Dividends and Stock Repurchases. Journal Of Financial Economics, 57(3), 355–84.

(46)

42

Appendix

Table 2 Family firms use of payout vehicles by percentile

Percentiles 10% 25% 50% 75% 90%

Dividends 0 0 .009116 .020115 .040127

Share repurchases 0 0 0 0 .002492

Table 3 Summary statistics of dividends(dividends/total assets) and share repurchases(share repurchases/total assets) in family firms by country

Summary Dividends Share repurchases

Country N Mean Std. dev Dev. Freq. Mean Std. dev. Dev. Freq.

Belgium 26 0.179 0.032 1,493 .001 .005 1,493 France 75 0.110 0.021 9,232 .001 .004 9,232 Germany 71 0.142 0.027 8,728 .000 .003 8,728 Italy 29 0.122 0.022 3,694 .001 .004 3,694 Netherlands 16 0.172 0.028 1,510 .001 .005 1,510 Portugal 13 0.147 0.028 765 .000 .003 765 Spain 34 0.151 0.029 2,108 .001 .005 2,108 Total 264 0.133 0.025 27,530 .001 .004 27,530

Table 4 Correlation table

the correlation between the two dependent variables (Dividends and Share repurchases (stated as buybacks) and the independent and control variables.

(47)

43

Table 5 Summary statistics of family firm size by country

Shareholder payout in family firms