CEO Compensation in Founder-Controlled Firms: A Comparison Between Family CEOs and Non-Family CEOs

When the founder is no longer the CEO of the firm he controls, he can entrust the position either to a family member or to a professional non-family CEO. Our study examines how founder-controlled firms compensate family CEOs versus non-family CEOs. Using regression analyses on a sample of large Canadian firms listed on the Toronto Stock Exchange (S&P/TSX) over a 7-year period (2002 to 2008), we find that the total compensation of family CEOs is lower than that of non-family CEOs, especially when the founder does not have excess voting rights. On the other hand, there is no difference in the incentive component of the compensation of family and non-family CEOs. Overall, our results tend to support the “optimal contracting” approach as opposed to the “rent extraction” approach in explaining founders’ behaviour in regard to CEO compensation when no excess-voting rights are involved.


Introduction
Most firms around the world, including large publicly traded firms, are controlled by families, often by the founder of the firm and/or members of the founder's family (Claessens et al., 2000;Faccio & Lang, 2002;La Porta et al., 1999).These families are particularly powerful as they often hold a significant share of the firm's voting rights and often sit on the board of directors (Bozec & Di Vito, 2019;Jaskiewicz et al., 2017).The governance literature often portrays founders as extraordinary beings whose personal qualities and active involvement in the firm make them an asset to the firm.Moreover, empirical results show that family firms perform better than non-family firms when the founder is present and is the largest shareholder and/or CEO (Anderson & Reeb, 2003;Villalonga & Amit, 2006;Miller et al., 2007, Fahlenbrach, 2009;Miller et al., 2011;Block, 2012).The beneficial contribution of the founder would even be observed when the latter no longer holds the position of CEO but continues to be a member of the firm's board of directors (Villalonga & Amit, 2006).
The objective of this study is to examine the role played by the founder on the governance of the firm, and specifically in elaborating the compensation policy of the CEO.When the founder is no longer the CEO of the firm he controls, he can entrust the position either to a family member (FMCEO) or to a professional, non-family CEO (NFMCEO).In this study, we examine whether the compensation offered to CEOs differ depending on who occupies the position.Is the founder more generous to a family member than to a professional CEO?In other words, in founder-controlled firms, how is the compensation of a CEO who is a member of the founder's family compared to that of a non-family CEO?Our study aims to precisely answer the latter question.
The financial literature on CEO compensation argues that CEO compensation can be used as a mechanism to alleviate agency conflicts between CEOs and shareholders.However, CEO compensation may also raise agency conflicts in founder and family-controlled firms.Accordingly, there are two competing theoretical approaches as to how family CEOs and non-family CEOs are compensated within founder-controlled firms.On the one hand, based on the "optimal contracting" approach (Gomez-Meja & Wiseman, 1997), because of the closer economic and emotional ties that family CEOs have with the firm, conflicts of interest are less severe and thus family CEOs are expected to receive less total compensation and/or less incentive-based compensation.On the other hand, based on the "managerial power" approach (Bebchuk & Fried, 2003), founders may use their power to extract private benefits by guiding compensation practices within the firm in a way that provides family CEOs with comparatively more generous compensation.As discussed in Bebchuk et al. (2002), this is a particular concern when the controlling shareholder holds a proportionately larger share of votes than equity (i.e., excess voting rights) because the incentives to extract private benefits are exacerbated.
We performed various regression analyses on a sample of large Canadian firms listed on the Toronto Stock Exchange (S&P/TSX) over a 7-year period (2002 to 2008).Our results reveal that the total compensation of family CEOs is lower than non-family CEOs, especially when the founder does not have excess voting rights.On the other hand, there is no difference in the incentive component of the compensation of family and non-family CEOs within founder-controlled firms.Overall, our results tend to support the "optimal contracting" approach.
This study complements previous studies on the compensation of CEOs of family firms.Some studies show that family CEOs are generally paid less than non-family CEOs (McConaughy, 2000;Gomez-Mejia et al., 2003).To our knowledge, the only studies that have looked at the specific case of founder-controlled firms are the studies by Li and Srinivasan (2011) and Jaskiewicz et al. (2017).However, these studies compare the compensation of non-family CEOs in founder-controlled firms with non-founder firms.Our study extends this research area by comparing the compensation of family and non-family CEOs within founder-controlled firms while controlling for the moderating effects of the founder's excess voting rights.

CEO Compensation in Family Firms
When a founder wishes to retire from the CEO position, he or she is faced with two CEO replacement choices: either to recruit a family member (FMCEO) or to recruit an individual with no family ties (NFMCEO).There are several assumptions about how family CEOs and non-family CEOs are paid in founder-controlled firms.Two distinct theoretical approaches arising from agency theory have been developed in the literature to explain CEO compensation.On the one hand, the "optimal contracting" approach considers compensation as a mechanism that the board of directors uses to align managers' interests with those of the shareholders and thus, mitigate the agency problem (Gomez-Meja & Wiseman, 1997).On the other hand, the "managerial power" approach considers compensation as a source of conflicts of interest rather than an alignment mechanism arguing that CEOs may have the power to influence their own compensation schemes (Bebchuk & Fried, 2003).
According to the optimal contracting approach, the total compensation of family CEOs is lower and less incentive-based than that of non-family CEOs of family firms.First, family CEOs are more attached to the firm, both economically and emotionally (Gomez-Mejia et al., 2003).Their interests are more naturally aligned with those of the firm, there is no need to offer them incentive compensation comparable to that of non-family CEOs.Second, the relationship between the family CEO and the founder is likely to be closer, which could alleviate problems of information asymmetry and reduce the level of compensation offered.Finally, family CEOs enjoy greater job security than non-family CEOs (Combs et al., 2010).Chen et al. (2013), for example, show that CEOs affiliated with the founder's family stay longer and are less likely to be replaced, even when their performance is unsatisfactory.Thus, family CEOs are probably willing to receive less total compensation in exchange for the additional job security (Gomez-Mejia et al., 2003).
In contrast, the managerial power approach predicts that the total compensation of family CEOs and/or the fixed component of compensation should be greater than that of non-family CEOs.Because of their status as founders, their in-depth knowledge of the firm and the significant share of voting rights they hold, founders have significant influence on decisions, including decisions relating to CEO compensation (Li & Srinivasan, 2011).In this context, founders can use their power to extract private benefits at the expense of minority shareholders, for example by granting excessive compensation to the CEO from their own family (Bebchuk et al., 2002).In doing so, founders would show altruism towards their family members (Schulze et al., 2002(Schulze et al., , 2003).Parents, for example, are naturally generous to their children, caring for them, protecting them, and favoring them (Eddleston & Kidwell, 2012).Parents are not only generous to their children because they love them, but also because they feel compelled to do so, as not being generous would even harm their own well-being (Lubatkin et al., 2005).
Prior empirical studies yield mixed results (See the study by Michiels et al. (2022) for a literature review).These studies mainly compared the compensation of CEOs of family firms and non-family firms.Some show that total CEO compensation is lower and/or less incentive in family firms (Baek & Fazio, 2015;Barontini & Bozzi, 2018;Combs et al., 2010;De Cesari et al., 2016;Pooser et al., 2017;Tinaikar, 2014;Yarram & Adapa, 2020).Others show the opposite (Basu et al., 2007;Cheng et al., 2015).In the more specific context of founder-controlled firms, Li and Srinivasan (2011) conclude that the total compensation of non-family CEOs is lower than that of CEOs of other firms, while Jaskiewicz et al. (2017) find it less incentive.
There is relatively less empirical evidence on the comparison of family and non-family CEOs within family firms.Some studies show that family CEOs are generally paid less than non-family CEOs (McConaughy, 2000;Gomez-Mejia et al., 2003).Others, on the other hand, show that in family firms, the compensation of family CEOs is higher than that of non-family CEOs (Cai et al., 2013;Jong & Ho, 2019;Kim & Han, 2018), in particular because their incentive compensation is higher or total compensation is more closely linked to performance (Cui et al., 2018;Kim & Han, 2018).Surprisingly, there are very few studies that have verified the impact of excess voting rights on CEO compensation.As discussed in Bebchuk et al. (2002), when the controlling family holds a proportionately larger share of votes than equity, it exacerbates conflicts of interest between the family and the rest of the shareholders and results in negative effects on the performance of family firms (Bozec & Di Vito, 2019;Claessens et al., 2002;Gompers et al., 2010;Villalonga & Amit, 2006).Masulis et al. (2009) and Amaoku-Adu et al. (2011) examined CEO compensation in dual-class firms, which have a significant proportion of family firms.Masulis et al. (2009) show that CEO compensation increases with the level of excess voting rights and that this positive relationship is even stronger for family CEOs.Amaoku-Adu et al. ( 2011) conclude that family CEO in dual class firms receive higher incentive-based compensation than non-family CEO in dual class firms.
The previous studies have provided insights into the influence of families on CEO compensation.However, given the heterogeneity of family firms, some important issues remain under-examined.First, previous studies do not distinguish family firms in which the founder is the major shareholder and/or board member from other firms, which we call founder firms.Previous studies generally include founder firms and descendant firms in family firms.The distinction is important when studying conflicts of interest between the family and the rest of the shareholders (Miller et al., 2007(Miller et al., , 2011;;Jaskiewicz et al., 2017).Second, the presence of excess voting rights, which is widely associated with family firms, has been only minimally controlled for in previous studies.It has not been considered at all in the specific case of founder-controlled firms, which is what we aim to do in this study.

Research Hypotheses
Given the theoretical arguments from the optimal contracting and managerial power approaches and the empirical evidence, it is difficult to predict how, within founder-controlled firms, the compensation of CEOs from the founder's family (family CEOs) compares to that of non-family CEOs.For these reasons, we formulate the following hypotheses in a null form: H1: In founder-controlled firms, the total and incentive-based compensation of family CEOs do not differ from those of non-family CEOs.
H2: In founder-controlled firms, when the founder has excess voting rights, the total and incentive-based compensation of family CEOs do not differ from those of non-family CEOs.

Sample
We have selected all Canadian firms listed on the S&P/TSX index for the years 2002 to 2008.From our initial sample, we eliminated financial sectors (n=175), observations with incomplete data (n = 167) and widely held firms (n=459).Furthermore, because our study aims to contrast CEO compensation between family CEOs and non-family CEOs within founder-controlled firms, we have removed firms for which the founder is the CEO (n=180).Our final sample is an unbalanced panel composed of 172 firms and 656 firm-year observations.

Founder-Controlled Firms and CEOs Identity
Using management proxies available on SEDAR, we have classified firms according to their ownership structure and the identity of their ultimate shareholder.To do this, we have followed the same methodology as La Porta et al. (1999), Claessens et al. (2000) and Faccio and Lang (2002).The ultimate shareholder is a shareholder who holds, directly or indirectly through another firm, the largest number of voting rights (at least 10%).Once the ultimate principal shareholder is identified, we verify whether it is the founder of the firm.To do so, we rely on firms' proxy circulars, but also on various Internet sources, such as the firms' corporate website.Founder-controlled firms are defined as controlled firms in which the ultimate principal shareholder is the original founder of the firm.
Our focus is on founder-controlled firms that are no longer run by the founder.Therefore, two possibilities can be observed regarding the type of CEO in those firms: 1) family CEO (FMCEO), when the CEO is not the founder but a member of the founder's family and 2) non-family CEO (NFMCEO) when the CEO is a professional manager without any relationship to the founder nor the members of the founder's family.We identified the identity of the CEOs using SEDAR, corporate website, but also from Bloomberg Business.

CEO Compensation and Control Variables
The CEO annual compensation data is manually collected and includes the various forms of compensation: 1) cash-based compensation (CASH) which includes the yearly base salary and the cash bonus; 2) long-term incentive-based compensation (INCENTIVE), which includes the number of yearly stock options and stocks granted to the CEO as compensation; 3) the CEO's total annual compensation (TOTAL), which is the sum of CASH, INCENTIVE, and indirect compensation such as pensions and benefits, and executive perquisites.We consider that the yearly stock option grants and share-based compensation are part of a long-term incentive plan established by the Board of Directors, tying CEO compensation to the market value of the firms' shares to encourage CEOs to maximize firm value.To determine the estimated value of the stock options granted, we followed Gomez et al. (2003) and multiplied by 25%, the exercise price of the options.
Following previous studies on CEO compensation (Combs et al., 2010;Amoako-Adu et al., 2011;Li & Srinivasan, 2011, among others), we control for factors that impact on CEO compensation.We gathered the following CEO characteristics: their age (CEO AGE), the number of years in office (CEO TENURE), if they are the president of the firm (CEO, PRESIDENT), and if they are the president of the Board of Directors (CEO CHAIRMAN).We also controlled for the CEO's percentage of voting rights (CEO OWNERSHIP).These features reflect the levels of CEO competence, power, and motivation (Combs et al., 2010).
From the StockGuide database, we also gathered the following firm-specific characteristics: firms' performance measured by the return on assets ratio (ROA), their size measured by total assets (SIZE), their age measured by the number of years in operation (FIRM AGE), their level of debt (DEBT), their growth opportunities (PRICE-TO-BOOK), their level of risk, measured by the level of R&D investment (R&D), their corporate governance quality (GOVERNANCE) as measured by the Globe and Mail's Report on Business (ROB) index, and the level of institutional ownership (INSTITUTIONAL).We also controlled for the founder's influence using the proportion of founder ownership (FOUNDER OWNERSHIP).

Empirical Model
The hypothesis tests are performed using cross-sectional time series regressions with fixed effects for years and the industrial sectors.Our main empirical model is as follows: Our dependent variables measuring CEO compensation (TOTAL, CASH and INCENTIVE) are standardized by using the log function.Our main variable of interest is the family CEOs (FMCEO).To directly compare family CEOs' compensation with that of non-family CEOs (our reference group) within founder-controlled firms, we include non-founder firms dummy (NON-FOUNDER FIRMS) in our model which comprises CEOs of non-founder-controlled firms.

Descriptive Statistics
The descriptive statistics of our sample is presented in Table 1.In column 1, we present the descriptive statistics of our full sample.The statistics presented in columns 2 and 3 refer only to founder-controlled firms, depending on whether the CEO is a member of the founder's family (FMCEO) or a professional CEO without any connection to the founder or the founder's family (NFMCEO).Next, we perform mean comparison T-tests for these two subsamples.
The average total compensation of CEOs in our sample of Canadian firms is 2,698,503.00$.The cash-based component averages 1,447,339.00$ while the long-term incentive-based component averages 1,103,044.00$.When comparing the average compensation of CEOs within founder-controlled firms, we can see that the non-family CEOs generally have higher wages than family CEOs, but the latter receive more cash-based compensation.While no statistically significant differences are observed in the average total and cash-based compensation between the two subgroups, FMCEOs incentive-based compensation is, on average, significantly lower than that of NFMCEOs.These results partially contradict our first research hypothesis as incentive-based compensation is significantly different between family CEOS and non-family professional CEOs.
Regarding the characteristics of founder-controlled firms, we find FMCEO firms to be, on average, significantly older, larger, and better governed than NFMCEO firms.On the other hand, NFMCEO firms are significantly less indebted and invest significantly more in R&D than FMCEO firms.Lastly, when comparing CEO characteristics within founder-controlled firms, we find that on average, family CEOs are significantly younger than non-family CEOs.FMCEOs have a significantly longer tenure, own significantly more shares of the firm, and occupy the Chairman position significantly more often than NFMCEOs.

Testing H1
The results of our regression analyses to test our first research hypothesis is presented in Table 2.The first regression is run on TOTAL, the dependant variable representing total CEO compensation.Our results indicate that family CEOs receive significantly less compensation than non-family CEOs (Coefficient -0.296 P-value < 0.05).These results are consistent with the optimal contracting approach and suggest that family CEOs, because of their closer economic and emotional ties to the firm and greater job security, accept lower total compensation than non-family CEOs.These findings are also consistent with the results of studies comparing the compensation of family CEOs and non-family CEOs within U.S. family firms (McConaughy, 2000;Gomez-Mejia et al., 2003).However, when examining the different forms of compensation in regressions 2 and 3, using INCENTIVE (for incentive-based compensation) and CASH (for cash-based compensation) as dependent variables, we do not find any significant differences in the form of compensation granted to family and non-family CEOs in founder-controlled firms.H1 is therefore only partially supported, as there is a significant difference in TOTAL compensation between FMCEOs and NFMCEOs yet no significant difference in incentive-based compensation between the two types of CEOs.
Looking at our control variables, we note that total CEO compensation is negatively associated to founder ownership rights, suggesting that the more ownership the founder has, the lower the total compensation of the CEO.This concentration of ownership would provide founders not only the ability but also the incentive to effectively monitor their CEOs.These results and are in line with the optimal contracting approach suggesting that, in the absence of information asymmetry, compensation tends to be lower.The NON-FOUNDER FIRMS variable has a significant negative coefficient for TOTAL and CASH regressions.This indicates that in non-founder-controlled firms', CEOs total and a well as cash-based compensation are generally lower than that of founder-controlled firms' non-family CEOs.Note.This table presents results from multivariate regression analyses with fixed effects of CEO identity on total CEO compensation and its components.Our final sample is an unbalanced panel composed of 656 firm-year observations.See Appendix I for the definition of the other variables.The value P-value is corrected to the heteroscedasticity (White, 1980).
Many of the firm-specific and CEO-specific control variables have a significant impact on CEO compensation, but the signs may vary according to compensation-type.For instance, our results show that the size of a firm, its growth opportunities as well as the quality of corporate governance have a significantly positive effect on CEO compensation as larger and high-growth firms who are well governed have the resources and understand the governance-related benefits of aligning CEOs' interests to those of shareholders by granting higher wages.CEO tenure is negatively associated with total and incentive-based compensation.This finding is counterintuitive as CEOs' salaries should increase over time.A possible explanation is that in concentrated ownership firms, where CEO alignment is not a crucial issue as information asymmetry is minimal, maybe CEOs that have been managing the firm for a long time, have attained the upper limit of their compensation and do not wish to change their comfortable position and undertake the risk of working somewhere else.On the other hands our results show a positive association between CEOs who also occupy the positions of either President or Chairman and incentive-based compensation as these latter positions are often compensated with long-term focused incentives.

Testing H2
To test our second research hypothesis which predicts the impact of founders' excess voting rights on family versus non-family CEOs, we split our FMCEO independent dummy variable into SEP FMCEO (founder firms with family CEOs for which the founder holds excess voting rights) and NOSEP FMCEO (founder firms with family CEOs for which the founder does not hold excess voting rights).The results of these regressions are presented in Table 3. Note.This table presents results from multivariate regression analyses with fixed effects of CEO identity on total CEO compensation and its components.Our final sample is an unbalanced panel composed of 656 firm-year observations.See Appendix I for the definition of the other variables.The value P-value is corrected to the heteroscedasticity (White, 1980).
Our results presented in Table 3 are very similar to those in Table 2. However contrary to our null hypothesis H2, our results show a significantly negative association between NOSEP FMCEO and TOTAL compensation.Hence, we find that there is a difference in founder behaviour when compensating the firm's CEO, but it is only when founders do not have excess voting rights that family CEOs are granted lower compensation that non-family CEOs.In the context of family firms, family members are concerned with preserving socio-emotional wealth, and in the absence of excess voting rights, where the founders' wealth is vested in the firm, founders and family members are well aligned to enhance their firm's value.Thus, founders' families may be willing to accept to better compensate external professional non-family CEOs to make sure there is a good alignment of interests with the family.Our findings are thus in line with the optimal contracting hypothesis, as in the context of minimal information asymmetry, founders will compensate family members less than non-family members.
Moreover, we do not find any significant association between SEP FMCEO and any of our dependent compensation variables.Thus, founders endowed with excess voting rights do not compensate family CEOs differently than non-family CEOs within founder-controlled firms.This may be contrary to the rent-extraction hypothesis, as it would be expected, from this point of view, for founders with excess voting rights to better compensate their family members, which would be beneficial to the family as opposed to non-family CEOs.Or perhaps, founders who have less ownership stakes but have full control over their firms externalize the cost of their suboptimal decisions to minority shareholders, and not compensating family CEOs more modestly than professional non-family CEOs is yet another way to do so.
The impact of firm specific and CEO-specific control variables on all three dependent variables in Table 3 are very similar to what we observed in Table 2.
When looking at our results globally, we are obliged to partially reject both of our null research hypotheses H1 and H2, as our study shows a clear difference between the compensation of family and non-family CEOs in founder-controlled firms.We only partially reject these hypotheses, as the difference observed is specifically for the total compensation while no significant effect is observed in the case of incentive-based compensation (H1).Differences are also observed between founders' behaviours regarding family versus non-family CEO compensation, depending on whether they hold excess voting rights (H2).Our study shows that, in a founder-controlled setting, family CEOs' total compensation is lesser than non-family CEOs, but only in the case where founders do not hold excess voting rights.Hence, in line with other studies showing that excess-voting rights can alter founders' behaviour and affect their decision-making, our findings show that excess voting rights is also a key determinant on founders' behaviour regarding the compensation of their successor CEOs.

Conclusion
In this study, we examined how founders in founder-controlled firms compensate their family members compared to non-family members in the CEO position.While many studies have examined CEO compensation in family firms, none to our knowledge examined the specific case of founder-controlled firms to examine how the firm's founder compensates the firm's CEO.Our study also examined if the founders' behaviour regarding CEO compensation differs when they hold excess voting rights.Because excess voting rights are known in the literature to exacerbate agency costs in concentrated ownership first, we show that excess-voting rights is a key variable that may significantly influence founders' behaviour when it comes to compensating family versus non-family CEOs.In our study, we show that founders compensate family CEOs less than non-family CEOs but only when they do not hold excess voting rights.These findings illustrate the importance of considering excess-voting rights when examining compensation in concentrated ownership and family firms.As many countries in the world are endowed with concentrated ownership firms, our study may be generalizable to such economic settings.
Our study also has a few limitations, firstly, our data is not very recent.As our data is hand-collected, and the purpose of the study is to observe founders' behaviour regarding compensation, we believe that more recent data should not alter our results.Second, our results must be interpreted with caution.Although we controlled for a multitude of factors associated with firms and CEOs, we cannot rule out the possibility that the associations between the presence of a controlling founder and the CEO compensation policies are due to other unobserved differences in firm or CEO characteristics.However, this limitation is not unique to us but concerns all empirical studies on the subject.Finally, it would be interesting to examine more precisely the components of CEO compensation, not only the cash-based and incentive-based components, but also indirect compensation such as pensions and benefits, and executive perquisites.The latter are, however, included in our dependent variable representing the total compensation of CEOs.All these limitations create excellent avenues for research.We encourage researchers to further examine CEO concentration in family firms, especially in the presence of excess voting rights, hoping that our present study paves the way for future research.

Table 1 .
Descriptive statistics and univariate tests Note.This table presents descriptive statistics of our sample composed of 656 firm-year observations.This table also presents statistics and the results of mean comparison T-tests between family CEOs (FMCEO) and non-family CEOs (NFMCEO) within founder-controlled firms.See Appendix I for the definition of the other variables.

Table 2 .
Compensation of Family CEOs vs. Non-Family CEOs

Table 3 .
Compensation of Family CEOs vs. Non-Family CEOs according to the founder's excess voting rights