The effect of the departure of a founder-CEO on firm
performance
Amsterdam Business School
Name Roxanne Schultz
Number 10248684
BSc in Business Economics Specialization Finance
Supervisor Ilko Naaborg Completion 1/7/2014
Abstract
Founder-CEOs have a positive effect on firm performance. This study investigates the effect of a founder-CEO departure on firm performance. This effect is determined by the difference in changes in measured performance between firms with a founder-CEO departure and a control group with firms without a founder-CEO departure. The sample consists of 310 publicly traded, U.S. firms during the years 1985 – 2013, including 62 firms with a founder-CEO departure. Firm performance is measured by the return on assets and stock return. This research finds that the departure of a founder-CEO does not have an effect on firm performance.
Table of Contents
1. Introduction 3
2. Literature review 4
2.1. Founder-CEOs vs. successor-CEOs and firm performance 4
2.2. Founder-CEO departures 6
3. Hypothesis, methodology and data 7
3.1. Hypothesis and methodology 7
3.2. Data and descriptive statistics 8
4. Empirical results 11
4.1. Empirical results 11
4.2. Robustness check 13
5. Conclusion and discussion 15
1. Introduction
In the past few years, the role of the CEO has become a major subject of discussion. The problem exists that managers are acting in the benefit of their own wealth. This problem is called the principal-agent problem. The agency theory states that the principals, the CEOs, are tended to act in their own interest instead of in the interest of the agent, i.e. the shareholders. However, there is evidence to believe that CEOs who also founded the company are an exception to the principal-agent problem. Founder-CEOs may be more concerned about the wealth of the company and may have more interest in the long-term profitability of the firm than successor-CEOs. Several studies have found evidence that corresponds to these notions about founder-CEOs. For example, Anderson and Reeb (2003) have found a positive relation between a founder status and firm performance compared to firms with a nonfounder-CEO, measured by an accounting based and a market measurement. Furthermore, Fahlenbrach (2009) has found a positive effect of founder-CEOs on the stock market performance of the firm.
The main purpose of this research is to investigate if the departure of a founder-CEO has an effect on firm performance. Previous research has found that a firms with a founder-CEO perform better than firms with a successor-CEO (Anderson & Reeb, 2003, Fahlenbrach, 2009). This study differs from earlier research by
investigating the effect on firm performance when a founder-CEO leaves the company and a nonfounder becomes the new CEO.
The sample consists of 310 U.S. firms during the years 1985 – 2013 and are collected from ExecuComp. To test the effect of a founder departure on firm
performance, two performance measures are used. The first is an accounting measure, return on assets (ROA). The second is a stock market measure, the return on
investment (ROI). There are three groups of firms: firms with the departure of a founder-CEO, firms with the departure of a successor-CEO and a control group with firms without a CEO departure. The performance measures and groups with firms are hand-collected. To investigate the effect of a CEO departure on firm performance, the average of the ROA and ROI of 3 years before and 3 years after a CEO departure year are determined. The difference in changes in measured performance between the group with a departing CEO and the control group is the effect of a CEO departure on firm performance.
The remainder of the thesis is organized as follows. Section 2 gives an overview of previous research and their findings. Section 3 discusses the hypotheses, methodology and data, and provides summary statistics. Section 4 presents the empirical results. Section 5 provides a summary and concludes the thesis.
2. Literature review
In existing finance literature, a distinction can be made between research into the effect of founder control on firm performance and the effect of a (founder-)CEO departure on firm performance. This section sets out conceptions and findings of earlier research on the relation between founder-CEOs and firm performance.
2.1. Founder-CEOs vs. successor-CEOs and firm performance
There are several notions about the effect of founder control on firm performance. The first is that having a founder as CEO is positively related to firm performance. Wasserman (2003) states that founder-CEOs often own a large percentage of the equity. There is no difference between ownership and control which means that the agency problem does not exist. According to Fahlenbrach (2009), a founder-CEO may have more affinity with the firm, may have more organization-specific skills and may prefer a different risk profile compared with a successor-CEO. Furthermore, the benefits of founding-family ownership are that the family may have strong incentives to solve the free-riders problem by monitoring the managers since their wealth is correlated with the welfare of the company (Demetz & Lehn, 1985). Anderson and Reeb (2003) also suggest that founding families are more long-term oriented than other shareholders, because they often maintain a long-term ownership.
On the other hand, there is also the idea that founding-family ownership has a negative effect on firm performance. Founding-family members may choose their own family as top management executives, despite the fact that an outside executive could have better specific skills and/or be more talented and capable to run the firm (Anderson & Reeb, 2003). DeAngelo and DeAngelo (2000) note that the family could induce poor stock price performance by desiring for special dividends.
Overall, previous research has produced consistent evidence on the effect of a founder-CEO on the performance of the company. The main outcome is that a
nonfounder-CEO. For example, Adams, Almeida and Ferreira (2009) investigated the relationship between founder-CEOs and firm performance. Their sample consisted of firms from the 1998 Fortune 500 and measured firm performance during the years 1992 – 1999. Firm performance was measured using an accounting measure, return on assets, and a market-based measure, Tobin’s q. To test the effect of a founder-CEO on these performance measures, a regression analysis is used including a dummy variable for a founder-CEO status and control variables for firm size, firm age and stock volatility. Adams et al. (2009) conclude that, on average, founder-CEOs increase firm performance and only leave the firm when the firm is doing well.
A similar effect on firm performance was found by He (2008). The sample included firms that went public during the years 1998 – 2002. He found that financial performance a few years after an IPO is higher when the firm is managed by a
founder-CEO, and that the likelihood a firm survives an IPO is greater when a firm is founder-managed than when a firm is managed by a successor-CEO. Also, He (2008) found that the corporate governance structure influences the effect on firm
performance. The financial performance of firms is even higher when the founder-CEO does not only have the position as founder-CEO but also as chairman of the board. Anderson and Reeb (2003) also used ROA as a profitability-based measure of performance and Tobin’s q as a market-based performance measure to investigate the effect of founding-family ownership on firm performance. To investigate this effect they used Standard & Poor’s 500 firms in the years 1992 – 1999. They found that founder-CEOs and founder descendents and/or family members as CEOs are positively related to accounting profitability compared with nonfamily firms. However, only founder-CEOs and successor-CEOs have a positive effect on the measured market performance, family members as CEO do not. Furthermore, Anderson and Reeb (2003) found that the effect of increasing family ownership on firm performance is first increasing and then decreasing.
Earlier research also investigated the relationship between founder-CEO status and performance on the stock market. Fahlenbrach (2009) constructed a sample 2327 large, U.S firms that were publicly traded from 1992 through 2002. Just as Adams et al. (2009) and Anderson and Reeb (2003), he used Tobin’s q to measure the effect of a founder-CEO on the firm’s market valuation. Additionally, he investigated the relation between founder-CEOs and stock market performance. He found that firms with a founder-CEO do not only have higher firm valuations, but also perform better
on the stock market than comparable firms. A possible explanation for this is that founder-CEOs prefer a different risk profile and tend to pursue active growth activities. Finally, Johnson, Magee, Nagarajan and Newman (1985) found that the death of an executive who is also the corporate founder has a positive reaction on the share price.
2.2. Founder-CEO departures
Adams et al. (2009) also investigated the endogenous character of founder-CEOs. Their results provide evidence that firm performance has a negative effect on the founder-CEO status. This means that good firm performance has a negative effect on the likelihood that a founder stays the CEO of the company. The negative effect of firm performance on the founder-CEO status was also found by Wasserman (2003). The CEO-turnover rate is higher for successful founder-CEOs.
What are the circumstances that may cause a founder-CEO to leave the company? In addition to the negative relation between firm performance and founder-CEO status, there are some factors that could influence the departure of a founder. Boeker and Karichalil (2002) provided evidence that the likelihood of a founder departure increases as new ventures increase in size, measured by the number of employees. Their sample consisted of semiconductor start-ups, founded from 1983 through 1999. Furthermore, they showed that for firms where the founder has a greater share in ownership, there are lower levels of founder departures. Finally, Boeker and Karichalil (2002) found a relation between individual characteristics of founders and the likelihood that they leave the company. The likelihood that a founder leaves the firm is smaller when a founder has a function within the firm concerning research and development responsibilities or when the founder is also the chief executive of the firm.
Previous research has found that there is a positive relation between CEOs and firm performance by comparing the performance of firms with a founder-CEO and firms with a nonfounder-founder-CEO. This indicates that companies with a founder who is also the chief executive of the firm are performing better than firms with a successor-CEO. Therefore it is interesting to investigate what the effect is on firm performance when a founder-CEO departs and a successor-CEO takes his or her place.
3. Hypothesis, methodology and data
Earlier research has found that that on average firms with a founder-CEO perform better than firms with a nonfounder-CEO. Possible explanations are that a founder has more interest in the long-term profitability of the firm and/or more specific skills which results in higher firm performance than firms with a successor-CEO. If this is true, the predicted effect of a founder-CEO departure is a decrease in firm
performance because of the change from a founder-CEO to a successor-CEO.
3.1. Hypothesis and methodology
In this study, firm performance is measured using two measurements, return on assets (ROA) and return on investment (ROI). Return on assets is defined as the ratio of earnings before interest, taxes, depreciation and amortization (EBITDA) to the book value of the total assets. Return on investment is the stock return based on the change in share price and possible paid out dividend. To test whether the departure of a founder-CEO and the departure of a successor-CEO has an effect on firm
performance, the average ROA and ROI of 3 years before and 3 years after the year a CEO departed are taken to determine the change in measured performance. These are referred to as pretest and posttest measurements. In addition to the two research groups (firms with the departure of a founder-CEO and firms with the departure of a successor-CEO), a control group is created with firms without the departure of a CEO within the selected time frame. To create similar groups, the firms with the departure of a successor-CEO and the control group are selected in such way that these groups match on time period and firm size with the founder-CEO departure group.
The change in firm performance of the two groups with a CEO departure was compared to the change in firm performance of the group without a CEO turnover. The difference in change in performance between a departing CEO group and group without a CEO turnover is the effect of the departure of the CEO on firm
performance. This research design is also known as a nonequivalent control group design with pretest and posttest (Shadish, Cook & Campbell, 2002).
The statistical linear model that is used to test whether the departure of a CEO has an effect on the change in firm performance is:
where Y is the change in the measured performance, FD is a dummy variable with F=1 if a founder-CEO departed and zero otherwise, NFD is a dummy variable with NFD=1 if a successor-CEO departed and zero otherwise. , and are the model’s intercept, regression coefficients and disturbance term, respectively. All parameters are estimated by applying OLS. Because pretest and posttest measurements are taken within the same firms, no control variables are needed to include in the model. For both the performance measures, the following null hypotheses will be tested:
Anderson and Reeb (2003) and Adams et al. (2009) found that a founder status has a significant effect on firm performance. Therefore it can be expected that the departure of a founder-CEO has a significant effect on the change in firm performance, a
change that is different from the change in the control group, resulting in However, since the overall effect on performance of a founder-CEO is positive
according to Adams et al. (2009), Anderson and Reeb (2003), and Fahlenbrach (2009) one would expect that the effect of a founder leaving the firm on firm performance is negative, resulting in
3.2. Data and descriptive statistics
The data is collected from three groups of firms during the years 1985 – 2013. The first group is the group with firms where a founder-CEO has left. Data including company names, CEO names and departure dates of the CEOs are collected from ExecuComp. Since ExecuComp does not contain information about whether a CEO is also the founder or a co-founder of the firm, the internet is used to check whether the first CEO of the company is also the founder. Only a founder or co-founder who was also CEO in the firm, is identified as a founder-CEO. Founder descendents or other family members are not included.
The second group includes firms with a non-founding CEO turnover. For each year a founder-CEO left the company, two firms are selected of which the CEO left in
that same year. By including this group, the effect of the departure of founder-CEO can be compared with the effect on firm performance by the departure of a successor-CEO.
The third group is the group with firms without the departure of a CEO. To make sure this group is comparable to the founder departure group, I selected for each firm with a founder departure the two firms with the closest amount of total assets in the year the founder-CEO left. In this way, the second group has on average
approximately the same firm size as the first group (see Table 1).
Table 1: Descriptive statistics
Variable Mean Min. Max. SD
Entire sample ROA1 13.61418 -26.3196 51.97843 10.05834 ROA2 11.81904 -52.62778 75.44835 11.87531 ROI1 1.506202 -6.38915 12.79431 2.270658 ROI2 1.175213 -5.858547 9.286994 2.099134 Total assets 6357.437 31.833 307717 22946.13 N 310 Firms with departure of founder-CEO ROA1 16.797 1.031811 51.97843 9.205099 ROA2 14.55556 -52.62778 50.16173 12.92493 ROI1 2.013693 -2.194858 8.124661 1.886619 ROI2 1.483546 -3.186933 5.752417 1.961418 Total assets 5501.754 31.888 113105 15893.96 ln(Total assets) 7.17 3.46 11.64 1.54 N 62 Firms with departure of a successor-CEO ROA1 14.20317 -22.45759 51.02229 10.54982 ROA2 12.33328 -46.86051 75.44835 11.14377 ROI1 1.28746 -6.110969 8.795408 2.28874 ROI2 1.168992 -5.688956 9.286994 2.151703 Total assets 7945.226 40.002 307717 30968.35
ln(Total assets) 7.16 3.69 12.64 1.69 N 124 Firms without departure of CEO ROA1 11.43379 -26.3196 33.97625 9.517842 ROA2 9.936529 -47.15135 41.93362 11.81659 ROI1 1.467357 -6.38915 12.79431 2.40347 ROI2 1.025964 -5.858547 7.876111 2.113233 Total assets 5197.489 31.833 115450 15301.8 ln(Total assets) 7.12 3.46 11.66 1.52 N 124
The entire sample consists of 310 firms with 62 firms with a departure of a founder-CEO, 124 firms with a departure of a successor-CEO and 124 firms without a departure of a CEO. ROA is the ratio of EBITDA to Total Assets. ROI is the monthly stock return based on change in share price and paid-out dividend. ROA1 is the average of the ROA 3 years before the year a CEO left, ROA2 is the average of the ROA 3 years after the year a CEO left. ROI1 is the average of the monthly stock return 36 months before the year a CEO left, ROI2 is the average of the monthly stock return 36 months after the year a CEO left.
To calculate the ROA for all the sample firms, the total assets and EBITDA of 3 years before the year a founder-CEO left and 3 years after that year are gathered from Compustat. I use EBITDA instead of net income in my research to be able to compare the results with previous research. The measured performance before the year a founder-CEO left is the average of the ROAs of the 3 years before that year and the same goes for the measured performance after the year a founder left.
Stock market performance is measured by the ROI. Monthly total returns are calculated by the change in share price plus possible paid out dividend divided by the old share price and are gathered from the Center for Research in Security Prices (CRSP). The measured change in stock market performance is the average of the monthly stock returns for the 36 months after the year a founder-CEO left minus the average of and 36 months before that year.
In Table 1 summary statistics of the entire sample, firms with a departure of a founder-CEO, firms with a departure of a successor-CEO and firms without a
of which a founder-CEO departed. The average change in ROA for the entire sample is -1.80% with a standard deviation of 8.94. The change in measured stock market performance of the entire sample has an average of -0.33% with a standard deviation of 3.21.
4. Empirical results
4.1. Empirical results
The main purpose of this research is to find out whether the departure of a CEO who is also the founder of the company has an effect on firm performance, which is
measured by return on assets (ROA) and return on investment (ROI). To investigate if the departure of a founder-CEO has an effect on firm performance, the change in ROA and ROI of firms with a founder departure, with a successor-CEO departure and without a CEO departure are compared. This effect on ROA and ROI is measured by calculating the change in the average ROA of 3 years before and 3 years after a CEO departure-year and comparing this change to the change in performance of the firms without a CEO departure. If the departure of a founder-CEO has an effect, the
expected outcome is that the change in performance of firms with a founder departure is larger than for firms without a CEO turnover. Furthermore, the sign of the effect of a founder-CEO departure is expected to be negative because of the positive relation between a founder status and firm performance which is found in previous research. A regression analysis is used to determine the effect of a founder-CEO status on firm performance and the significance of these effects. The main regression model assumptions were evaluated, including the assumptions of homoscedasticity,
independence of residuals and the normal distribution of the residuals. No significant violations or influential observations were found.
First, the change in ROA was inspected for the three groups. As can be seen in Table 2, the change scores of ROA for the three groups are all negative. This means that the average ROA for all firms is decreasing in the measured time period. Table 2 further shows that the firms with the departure of a founder-CEO experience the largest change in return on assets, followed by the firms with the departure of a successor-CEO. The differences between the change scores of ROA for the two groups with a CEO departure and the group without a CEO departure are presented in Table 3. The change in ROA for firms with a founder-CEO departure is 0.74% larger
than for firms without a CEO departure. The change in ROA of firms with a
successor-CEO departure is 0.37% larger than in the control group. This corresponds to the predicted outcome that a founder-CEO departure causes a larger change in measured performance. However, this effect of 0.74% is not statistically significant, t(309) = -0.53, p = 0.594. The corresponding 95% confidence interval (CI) is -3.49% to 2.00%. The effect of 0.37% is also not significant, t(309) = -0.33, p = 0.744 (95% CI is -2.61% to 1.86%).
Table 2: Pretest and posttest measurements, and change in ROA for each group (in %)
ROA1 ROA2 Change SE
Firms with departure of founder-CEO 16.80 14.56 -2.24 1.26 Firms with departure of successor-CEO 14.20 12.33 -1.87 0.80 Firms without departure of CEO (control group) 11.43 9.94 -1.49 0.76
Table 3: Mean difference in change in ROA between groups (in %) No departure of CEO SE
Departure of founder-CEO -0.74 1.39
Departure of successor-CEO -0.37 1.14
Table 4 presents the change scores of ROI. Just as with the change scores of ROA, these change scores are negative for all the groups. The change in stock return is the largest for the firms with the departure of founder-CEO, namely -0.53%. The smallest change in the measured performance has occurred in firms with the departure of a successor-CEO. Table 5 shows the mean differences between the change scores of ROI of the three groups. As can be seen, the change in return on investment for
firms with a founder-CEO departure is 0.09% larger than for firms without a CEO departure, where the average ROI is decreasing for all firms in the measured time period. This effect is however not statistically significant, t(306) = -0.18, p = 0.859 (95% CI is -1.07% to 0.90%). The change in ROI of firm with the departure of a successor-CEO is 0.32% smaller compared to the groups with firms without a CEO turnover. This effect is also not significant, t(306) = 0.79, p = 0.432 (95% CI is -0.48% to 1.13%).
Table 4: Pretest and posttest measurements, and change in ROI for each group (in %)
ROI1 ROI2 Change SE
Firms with departure of founder-CEO 2.01 1.48 -0.53 0.36 Firms with departure of successor-CEO 1.29 1.17 -0.12 0.29 Firms without departure of CEO (control group) 1.47 1.03 -0.44 0.31
Table 5: Mean difference in change in ROI between groups (in %) No departure of CEO SE
Departure of founder-CEO -0.09 0.50
Departure of successor-CEO 0.32 0.41
4.2. Robustness check
To check whether the results are robust to alternative specifications, the same
regression analysis is applied but with robust standard errors. The coefficients stay the same but standard errors are corrected for possible heteroscedasticity, which would mean that the errors do not have the same variance across all observations. As can be seen in Table 6 and 7, applying the regression analysis with robust standard errors
does not change the outcome: the effects of departure of founder-CEOs and successor-CEOs on the change in ROA and ROI remain non-significant.
Table 6: Mean difference in change in ROA between groups (in %) with robust standard errors No departure of CEO SE t Departure of founder-CEO -0.74 1.46 -0.51 Departure of successor-CEO -0.37 1.11 -0.34
Table 7: Mean difference in change in ROI between groups (in %) with robust standard errors No departure of CEO SE t Departure of founder-CEO -0.09 0.47 -0.19 Departure of successor-CEO 0.32 0.42 0.77
An alternative approach is used to check whether the outcome changes when the performance is measured 1 year before and 1 year after the year that a founder-CEO departed instead of the average change of 3 years before and after the departure year. These results could be more affected by time specific circumstances but
therefore also more sensitive to a CEO turnover. The results are presented in Tables 8 and 9. Table 8 shows the mean difference in change in ROA between firms with the departure of a founder-CEO, the departure of a successor-CEO and without the
departure of a CEO. The negative change in ROA of 1 year before and 1 year after the year with a founder-CEO departure is 1.79% larger than for firms without a CEO departure. This effect of 1.79% is not significant, t(309) = -1.28, p = 0.203 (95% CI is -4.56% to 0.97%). The mean difference in change in ROI is shown in Table 9. As can be seen, the negative change in ROA in firms with a founder departure is 1.28% larger than for firms without a CEO departure. This effect of 1.28% is also not
significant, t(306) = -1.32, p = 0.189 (95% CI is -3.19% to 0.63%). Since both effects of a founder departure on the change in firm performance are not significant, the outcome does not change.
Table 8: Mean difference in change in ROA between sample groups (in %), 1 year before and after CEO departure-year
No departure of CEO SE
Departure of founder-CEO -1.79 1.41
Departure of successor-CEO 0.73 1.15
Table 9: Mean difference in change in ROI between sample groups (in %), 1 year before and after CEO departure-year
No departure of CEO SE
Departure of founder-CEO -1.28 0.97
Departure of successor-CEO 0.74 0.80
5. Conclusion and discussion
This research is conducted to investigate whether there is an effect of the departure of a founder-CEO on firm performance. Firm performance is measured in two ways. The first measurement is an accounting based measurement: return on assets (ROA). The second measurement is an stock market based measurement: return on investment (ROI). To determine if a founder-CEO departure has an effect on the performance of the company, the change in ROA en ROI is determined. The same is done for two other groups: firms with the departure of a successor-CEO and control firms without a CEO turnover. The changes in measured performance of firms with the departure of a founder-CEO and the departure of a successor-CEO are compared with the changes in measured performance of firms without a CEO departure.
Results show that for all firms the ROA and ROI is decreasing, in which there is no statistical evidence that the change is larger for firms with a departed founder-CEO or with a departed successor-founder-CEO. Therefore it cannot be concluded that the departure of a founder-CEO has an effect on the profitability and the stock market performance of the firm.
Based on the positive relation that previous research has found between a founder status and firm performance, the predicted outcome is that firm performance deteriorates when the founder-CEO departs and a successor-CEO takes his place. The
results of this research are not in line with this prediction. This study finds that the switch from a founder-CEO to a successor-CEO does not have an effect on firm performance.
There are a couple of limitations in this research. Unknown factors causing a founder-CEO to depart are not included in this research. However, different factors causing a departure of a founder-CEO could have a differential effect on firm performance. At this point, it is unknown which causes have a positive and which causes have a negative effect on firm performance. Furthermore, the departure of a founder as a CEO does not necessarily mean that he or she completely leaves the company. The founder can still have an (large) equity stake and/or play an important role in the decision making of the firm. These conditions are not included in this research, but it may be interesting for future research to investigate if a founder completely leaving the company does have a significant effect on firm performance.
References
Adams, R., Almeida, H., and Ferreira, D. (2009). Understanding the relationship between founder-CEOs and firm performance. Journal of Empirical Finance, 16, 136-150.
Anderson, R. C. and Reeb, D. M. (2003). Founding-Family Ownership and Firm Performance: Evidence from the S&P500. The Journal of Finance, 58(3), 1301-1328.
Boeker, W. and Karichalil, R. (2002). Entrepreneurial Transitions: Factors
Influencing Founder Departure. The Academy of Management Journal, 45(4), 818-826.
DeAngelo, H. and DeAngelo, L. (2000). Controlling stockholders and the disciplinary role of corporate payout policy: A study of the Times Mirror Company.
Journal of Financial Economics, 53, 153-207.
Demsetz, H. and Lehn, K. (1985). The structure of corporate ownership: Causes and consequences. Journal of Political Economy, 93, 1155-1177.
Fahlenbrach, R. (2009). Founder-CEOs, Investment Decisions, and Stock Market Performance. Journal of Financial and Quantitative Analysis, 44(2), 439-466. Johnson, W.B., Magee, R.P., Nagarajan, N.J. and Newman, H.A. (1985). An analysis of the stock price reaction to sudden executive deaths. Journal of Accounting
and Economics, 7, 151-174.
Lerong, He. (2008). Do founders matter? A study of executive compensation, governance structure and firm performance. Journal of Business Venturing,
23, 257-279.
Shadish, W. R., Cook, T. D., and Campbell, D. T. (2002). Experimental and Quasi-
Experimental Designs for Generalized Causal Inference. Boston, New York:
Houghton Mifflin Company.
Wasserman, N. (2003). Founder-CEO Succession and the Paradox of Entrepreneurial Success. Organization Science, 14(2), 149-172.
