CEO turnover announcement: The Sarbanes-Oxley effect
Name: Mayra Werges
Student number: 10180796
Specialization: Finance and Organization
Field: Finance
Supervision: Mark Dijkstra
Abstract
This thesis investigates whether the Sarbanes-Oxley Act (SOX) of July 2002 has an impact on shareholder behavior, by analyzing the effect of a CEO turnover announcement on share price volatility. The sample used in this thesis consists of 123 CEO turnovers from S&P 500 listed firms in the periods 1997 until 2001 and 2003 until 2007. Percentage changes in share price volatility are calculated, in a one-year and a one-month event study, to examine the effect of a CEO turnover announcement. No statistical evidence is found for differences in share price volatility changes after a CEO turnover announcement in the pre- and post-SOX period. The one-year event study finds a robust significant negative effect on the share price volatility after a forced CEO turnover announcement. However in the same study, an economically significant positive effect on share price volatility changes is found for forced turnovers in the post-SOX period.
1. Introduction
In the history of a firm, a Chief Executive Officer (CEO) turnover is a major event. A CEO is responsible for maximizing shareholders’ value by taking management decisions that affects the profitability of a firm (Kind & Schläpfer, 2011). The ability, operating policies,
preferences and decisions of the new CEO can be different from the departing CEO, leading to uncertainty about future firm performance (Clayton et al., 2005). After a CEO turnover announcement, cumulative abnormal stock returns and share price volatility changes affects shareholders’ wealth.
The Sarbanes-Oxley Act (SOX) was introduced in July 2002. The act aims to increase the amount of monitoring, prevent management and accounting misconduct and to improve corporate governance by requiring more supervision, increase penalties for managerial misbehavior and increase board independence (Dah et al., 2014; Zhang, 2007). According to Hochberg et al. (2009) investors believe that the SOX will improve the transparency and disclosure requirements of a firm. This study will examine the effect of the SOX on shareholders’ behavior by analyzing the effect of a CEO turnover announcement on share price volatility. In addition, this study will research whether the effect of the SOX is greater when the announcement concerns a forced CEO turnover. The research question is formulated as followed: Is there a difference in the share price volatility change after a CEO turnover announcement before and after the introduction of the Sarbanes-Oxley Act?
The volatility approach used by Clayton et al. (2005) and Ting (2013) is applied by calculating the percentage change of the share price volatility before and after a CEO turnover announcement. The sample consists of 150 firms randomly selected from the S&P 500 1. An event study of a year and an event study of a month will be performed to analyze the effect on a short-time period and a long-time period. To estimate the effect of the SOX, each event study is divided into a pre-SOX period (1997-2001) and a post-SOX period (2003-2007).
In the one-year event study, a robust significant negative effect on the volatility change is found after a forced CEO turnover announcement. No significant difference between the volatility change after a forced and a voluntary turnover announcement is found in the one-month event study. In both event studies, an insignificant negative effect of the SOX on the volatility change is found. A statistically insignificant but economically
significant positive effect of the SOX is found for forced CEO turnover announcements in the post-SOX period for the one-year event study. In the one-month event study, no unambiguous
1 This index consist of the 500 largest firms in the U.S. as measured by their market capitalization and is compiled by the credit rating agency Standard & Poor's.
significant effect of forced turnovers in the post-SOX period is found. In conclusion, no statistical evidence is found for an impact of the SOX on the volatility change after a CEO turnover announcement.
The literature and empirical review are discussed in chapter 2 and 3. In chapter 4 and 5 the methodology and data are described. Chapter 6 presents the results, after which these will be discussed in chapter 7. Finally, in chapter 8 the conclusion is made. The references and the appendix can be found in chapter 9 and 10.
2. Theoretical Review
2.1 CEO turnover
La porta et al. (2000) defines corporate governance as the shareholder protection against self-serving behavior by employees of a firm, formulated into a set of regulatory, legal and judicial mechanisms. Shareholders need to be protected because of the agency problem. The agency theory can be described as a contract between the agent and the principal in which the agent (the board) makes decisions and performs tasks on behalf of the principal (the shareholders) (Jensen & Meckling, 1976). According to this theory, CEOs may possess goals that diverge from those of the shareholders. Therefore, the interests of the agent are not always perfectly aligned with the interests of the principal. From principals’ perspective, it is not always possible for the principal to ensure that the agent will make the optimal decisions without making any monitoring costs2. A CEO will engage in self-serving activities at shareholders’ expense due to the separation of ownership and control (Jensen & Meckling, 1976). This is called the agency problem.
Coughlan and Schmidt (1985) state that separation of ownership and control is a misleading formulation of the origin of the agency problem. Ownership is not always completely separated, top managers sometimes own a small fraction of a firms’ equity. And even if ownership is completely separated from control, no agency problem exists if all information is available to the shareholders at no cost. Instead, Coughlan and Schmidt (1985) think it is better to say that the agency problem is a result of asymmetric information caused by the delegation of authority from the principal to the agent.
The identification and termination of a poorly performing CEO, a forced turnover, is an internal corporate governance control mechanism (Defond & Hung, 2004). For example, shareholders have the ability to use their voting rights to fire a low quality CEO and the board
of directors have the right to dismiss a CEO (Kind & Schläpfer, 2011). This will align the objectives of the shareholders with those of the CEO. If the CEO does not comply the
shareholders’ objectives, the CEO risks to lose his/her job. These internal control mechanisms will not always reduce or eliminate the agency problem. According to Chemmanur and Fedaseyeu (2012), an incapable CEO will sometimes remain because the board faces a
coordination problem. A coordination problem arises from the fact that every director receives a private signal about the quality of the functioning of the CEO and has no knowledge of the signals of the other directors (Chemmanur & Fedaseyeu, 2012). Even though the board discusses the quality of the CEO, there is a chance that a poorly performing CEO will not be fired when there are not enough votes against the current CEO. Even when a majority of the board of directors privately indicate that the CEO is performing poorly. If directors do not know the private signals of the other directors it is harder to coordinate the decision of the board. This problem is more likely to occur as the board becomes larger (Chemmanur & Fedaseyeu, 2012).
2.2 Abnormal stock returns and share price volatility
According to Bonnier and Bruner (1989), the abnormal return of a management change announcement can be attributed to an information effect and to the real effect. The information effect is negative when the change suggests that the firm’s performance was worse than the market initially thought (Bonnier & Bruner, 1989). The real effect of a CEO change is positive because it is in the interest of the shareholders. Kind and Schläpfer (2011) believe that shareholders should interpret a CEO turnover as a positive signal because a CEO will be fired when his/her qualities are poor and the expected costs of the dismissal will be lower than the expected returns flowing from the CEO departure. The internal corporate control hypothesis states that shareholders will benefit from a management change following poor firm performance (Bonnier & Bruner, 1989). Bonnier and Bruner (1989) argue that there will be a wide range of positive and negative abnormal cumulative returns, depending on the magnitude of the negative information effect compared to the positive real effect.
Clayton et al. (2005) developed three hypotheses to identify the impact of a CEO turnover on the share price volatility. The ability and strategy hypothesis predicts an increase in share price volatility due to an increase in uncertainty about the management’s ability to run the firm and the firm’s strategic direction after a CEO turnover. According to the
scapegoat hypothesis, all CEO’s have equal abilities and use them at the same optimal level. The board of directors will conduct a policy of firing CEOs after poor firm performance. A
CEO can be seen as a scapegoat, since dismissal occurs due to random factors that result in poor firm performance, rather than the lack of ability or effort level of the CEO. After controlling for pre-event volatility and firm performance changes, the scapegoat hypothesis states that a CEO turnover announcement has no direct effect on the share price volatility.
A change of a CEO can directly affect the investment and financial policy, the organizational strategy and the performance of a firm (Bertrand & Schoar, 2003). This may increase uncertainty about future firm performance, strategy, solvability and the management quality. The market may revise its expectations about firm value more often or more intensely than in the past because of an increased shareholders’ sensitivity to new information about the strategy and ability of the new CEO (Clayton et al., 2005). This will increase the share price volatility. Tests for shifts in the volatility of share prices, also known as the volatility
approach, can be used to evaluate the impact of a CEO turnover (Warner et al., 1988). 2.2 Sarbanes-Oxley Act
In response to a number of corporate scandals, like Enron and World-com, the U.S. Congress introduced the Sarbanes-Oxley Act in July 2002 (Zhang, 2007). The intent of the SOX is to restore shareholder confidence through greater transparency and additional corporate
disclosure requirements (Filbeck et al., 2011). The main objectives of the SOX, carried out by the Public Company Oversight Board, are (Filbeck et al., 2011):
- to supervise and register public accounting firms;
- to establish standards related to auditing and internal control;
- to increase auditor independence by restricting accounting firms to provide both auditing and non-auditing services to a firm;
- to improve corporate responsibility by stimulating auditing committee independence, executive certification for the accuracy of financial reports and penalties for board members related to financial restatements;
- to enhance financial disclosure, prohibiting personal loans to executives, provide a code of ethics and oblige the disclosure if the audit committee includes a financial expert;
- to increase the penalties for board members who have been committing in corporate fraud. Corporate governance is more effective if law enforcement institutions are capable to protect shareholders from self-interested CEOs that abuse their position and harm shareholder value (Defond & Hung, 2004). Defond and Hung (2004) find that strong law enforcement3
3 Strong law enforcement is measured by an index that captures aspects like the magnitude of government corruption and the effectivity of the judiciary.
institutions are related to improved CEO turnover-performance sensitivity, that is the possibility of a CEO turnover is higher when firm performance is poor. Therefore, these institutions are able to protect shareholders’ property rights. However, Kaplan and Minton (2012) believe this increased sensitivity is not caused by the legislation of the SOX but that it is related to increases in block holdings. Blockholders4 have greater incentives and capacities to monitor a CEO because the relative monitoring costs, compared to their equity value, are smaller than for shareholders with a small fraction of firm’s equity (Kaplan & Minton, 2012). Board independence also increases CEO turnover-performance sensitivity, Kaplan and Minton (2012) find a statistically significant negative relation between non-independent boards and turnover-performance sensitivity.
Davidson et al. (2010) find that after the introduction of the SOX, forced CEO
turnover is less likely and CEO tenure is not significantly shortened. This can be explained by the fact that they found an increase in the risk aversion5 of CEOs in the post-SOX period due to an increased monitoring level by the board and financial press. As a result of the increase in risk aversion, CEOs are less likely to take risks and, therefore, less likely to engage in self-serving activities that, in turn, increases the risk of losing their job. For the Fortune 5006 the total number of CEO turnovers decreased from 19.15% in the period 1997 to 2002, to 14.70% in the period 2003 to 2007 (Kaplan & Minton, 2012). Hazarika et al. (2012) find an
insignificant positive relation between earnings management7 and the likelihood of a forced turnover after the introduction of the SOX. The study of Linck et al. (2009) examines the effect of the SOX on board composition. In the post-SOX period compared to the pre-SOX period, there is an increase in the number of lawyers, consultants, financial experts and retired executives in the board of directors. Link et al. (2009) find a significantly positive relationship between board size and the SOX and a significantly positive relationship between board independence and the SOX. The audit committees met twice as many times a year (Link et al., 2009). In conclusion, the SOX affects board composition and the behavior of the CEO.
There are two different views regarding the possible impact of the SOX on
shareholders (Hochberg et al., 2009). Proponents of the SOX say that the reduction in agency costs due to improved transparency and disclosure requirements are greater than the costs of
4 A blockholder owns a large part of firm’s equity.
5 Risk aversion is measured by the average net risky investments divided by total assets.
6 The Fortune 500 is a ranking of companies from the United States, ordered from largest to smallest, based on their annual turnover.
7 Earnings management is measured by dividing the performance-adjusted absolute abnormal discretionary accruals by the assets of a firm.
compliance of the SOX. Opponents believe that the costs of compliance of the SOX will be higher than the agency benefits. However, Hochberg et al. (2009) does not mention who the proponents and the opponents exactly are. Hochberg et al. (2009) find statistical evidence for the positive view of the SOX by shareholders. They suggest that investors expect that the SOX will better align the interests of the CEO with those of the shareholders, and therefore will increase shareholders’ value.
3. Empirical Review
3.1 Abnormal stock returns
Existing literature find a positive share price reaction to a CEO turnover announcement. Using 367 turnovers of New York Stock Exchange (NYSE) listed firms between 1974 and 1983, Weisbach (1988) indicates a significant cumulative abnormal return of 0.35% in a day event window. Ting (2013) supports these findings with a positive significant three-day cumulative abnormal return of 0.19% and a five-three-day cumulative abnormal return of 0.23%, using a sample of firms listed on the TWSE (Taiwan Stock Exchange) from 2002 until 2006. The study of Huson et al. (2004) uses a sample of firms listed in the Forbes annual compensation survey8 between 1971 and 1995. It finds a cumulative abnormal return of 0.34% for the full sample. Different cumulative abnormal returns are found for forced turnovers and voluntary turnovers. The average cumulative abnormal return is higher for forced turnovers, namely 2.15%, compared to voluntary turnovers, namely 0.26%. Bonnier and Bruner (1989) find a significantly positive cumulative abnormal two-day return of 2.48% after the announcement of a change in senior management of distressed9 firms in the New York or American stock exchange in the period from 1969 until 1983. This is consistent with the internal corporate control hypothesis, which states that shareholders will benefit from a management change following poor firm performance. The study of Adams and Mansi (2009) looks at the effect of a turnover on total firm value, including publicly traded debt and equity, in the period from 1973 until 2000 using the database of Huson et al. (2001). They find wealth transfers from bondholders to shareholders. By way of illustration, forced CEO turnover announcements result in a cumulative abnormal return of approximately -0.73% to bondholders and 2.43% to shareholders.
8 The Forbes executive compensation survey consist of the 800 largest firms in the U.S.
9 A distressed firm is a firm that is underperforming financially. Underperforming is defined as negative earnings in the last quarterly report before the turnover. Dividend no earlier than 24 months is eliminated and no
However, existing literature find also a negative share price reaction to a CEO
turnover announcement. Using a database of firms listed in the UK on the All Share Index in the period from 1990 until 1995, Dedman and Lin (2002) find a negative market reaction to announcements of top executive departures. Particularly, an average cumulative abnormal return of -0.72% on the day of the announcement and an average cumulative abnormal return of -1.04% on a three-day event window. This negative reaction is higher when a CEO is forced out, namely -3.98% on the day of the announcement and -3.40% on a three-day event window. In contrast to Bonnier and Bruner (1989), Khanna and Poulsen (1995) find a negative announcement effect of –0.96% in a five days before to one day after the
announcement event study with a sample of distressed firms only in the U.S.10 from 1980 until 1990. A negative effect suggests that a managerial turnover is not seen as a remedy for financial distressed firms, but it worsens the situation. Therefore, the study of Khanna and Poulsen (1995) rejects the internal corporate control hypothesis.
Summarized, the effect of a CEO turnover announcement on the share price performance can be either positive or negative. It can be concluded that it is difficult to distinguish the real effect from the information effect. Like Bonnier and Bruner (1989) argue that there is a wide range of positive or negative abnormal cumulative stock returns. Even when the sample only consists of distressed firms to minimize the information effect, because it is expected that the market already recognized the poor firm performance, the studies of Bonnier and Bruner (1989) and Khanna and Poulsen (1995) come to opposite conclusions. 3.2 Share price volatility
There are two studies that examine share price volatility changes after a CEO turnover announcement, Clayton et al. (2005) and Ting (2013). Clayton et al. (2005) find, in a sample of 872 turnovers from the Forbes annual compensation survey in the period from 1979 until 1995, that the share price volatility increases after a CEO turnover announcement. This increase is larger for forced turnovers, approximately 24 %, than for voluntary turnovers, approximately 10%. Huson et al. (2004) give a reason for the smaller changes in volatility after a voluntary turnover. In contrast to a forced turnover, a voluntary departure does not have to be associated with poor performance of the CEO. The voluntarily leaving CEO is not necessarily of poor quality (Huson et al., 2004). Even though the board will appoint the candidate with the highest expected quality to make the voluntarily turnover a success, it does
10 Khanna and Poulsen (1995) obtained a list from of all public firms filing for bankruptcy protection provided from the U.S. Securities and Exchange Commission (SEC).
not mean that the new CEO has better qualities than the departing CEO.
Ting (2013) finds a decrease in volatility after a CEO turnover instead of an increase, based on a sample of companies listed on the TWSE from 2002 until 2006. In an event study of 31 to 90 trading days before and after a CEO turnover, the overall volatility decreases with 8.93%. When using a one-year event window, the overall volatility decreases with 7.36%. The decrease in volatility is the highest when the successor has less power11 than the former CEO, namely 17.6% (Ting, 2013). Ting (2013) argues that when the new CEO has less power to influence decisions of the board, these decisions will reflect the opinion of the new CEO to a lesser extent. The risk of judgment errors, i.e. differences in the decision preference of the executives, is smaller in boards where decisions are made by a group of executives compared to a firm in which only the CEO dominates the decisions. Smaller risks of judgment errors will decrease share price volatility (Ting, 2013). Concluding, the results of Clayton et al. (2005) and Ting (2013) give opposite directions of the volatility changes.
4. Methodology
The research question is formulated as followed: Is there a difference in the share price volatility change after a CEO turnover announcement before and after the introduction of the Sarbanes-Oxley Act? To answer the research question the following regression is used: Ln(volatilityt+1/volatilityt-1)i,t = β0 + β1 Dum Forcedi,t + β2 Dum SOXj + β3 Dum SOXj * Dum
Forcedi,t + β4 CEO agei,t + β5 Dum CEO genderi,t + β6 Dum Founderi,t + β7 CEO Tenurei,t + β8 Ln(S&P 500 Volatilityt+1/S&P 500 Volatilityt-1)i,t + β9 Pre-turnover NOI/assetsi,t + β10 market-to-book ratioi,t + β11 Ln(Mrkt. Cap.)i,t + β12-20 Dum S&P500 sectori,t + ε
The outcome of this regression can determine whether the SOX affects share price behavior after a CEO turnover announcement and whether this effect is greater when it concerns a forced turnover. To test if the outcome of Clayton et al. (2005), namely an increased volatility after a forced CEO turnover compared to a voluntary turnover, is feasible for the sample used in this study, hypothesis one is formulated as followed:
Hypothesis 1: The volatility change for a forced turnover is larger than the volatility change for a voluntary turnover after a CEO turnover announcement.
A forced dummy is included in the regression to estimate the effect of a forced turnover on
11 Power is measured by a power index that contains four CEO power dimensions: expert power, prestige power, ownership power and structural power.
the volatility change. The dummy is 1 if the CEO turnover is forced and 0 if the CEO turnover is voluntary.
To examine the impact of the SOX on the behavior of shareholders, the effect of the SOX on the market reaction after a CEO turnover announcement is investigated. On top of this, this study will investigate whether this effect is greater when the announcement concerns a forced turnover. So, the second and third hypotheses are formulated as followed:
Hypothesis 2: The SOX has an impact, negative or positive, on the volatility change after a CEO turnover announcement in the post-SOX period.
A SOX dummy is added in the regression. The dummy is 1 if the CEO turnover is in the post-SOX period and 0 if the CEO turnover is in the pre-post-SOX period.
Hypothesis 3: The impact of the SOX on the volatility change is larger when the turnover is forced.
To investigate this hypothesis, an interaction between the forced dummy and the SOX dummy is included in the regression.
The three hypotheses will be tested the following way: Hypothesis 1: H0: β1 = 0 H1: β1 > 0
Hypothesis 2: H0: β2 = 0 H1: β2 ≠ 0
Hypothesis 3: H0: β3 = 0 H1: β3 ≠ 0
The hypotheses will be tested with a t-test. Robust standard errors are used to control for the possibility of heteroskedasticity12. β
1 is significantly13 larger than zero when the t-value is larger than 1.65714. β
2 is significantly different from zero when the t-value is larger than 1.980 or smaller than -1.98015. β3 is significantly different from zero when the t-value is larger than 2.056 or smaller than -2.05616.
To follow the studies of Ting (2013) and Clayton et al. (2005), the control variables Ln(Firm Size), market-to-book ratio and Pre-turnover net operating income (NOI)/assets are included in the regression. Huson et al. (2001) find a positive relation between firm size and the probability of a CEO turnover. Ting (2013) measures firm size by taking the equity market value at the end of the fiscal year. This is the market capitalization of a firm, to be
12 Heteroskedasticity means that the variance and standard deviation of the standard errors are not constant over the entire graph of the sample data.
13 Significantly means a p-value of 0.05 or smaller. Meaning that the change of rejecting H
0 when H0 is true, in the long run, will be 5% or less of the time.
14 The critical t-value of a one-sided t-test at a significance level of 5% and 122 degrees of freedom. 15 The critical t-value of a two-sided t-test at a significance level of 5% and 122 degrees of freedom. 16 The critical t-value of a two-sided t-test at a significance level of 5% and 26 degrees of freedom.
precise, Ln(Firm size) is named Ln(Mrkt. Cap.). The market-to-book ratio is measured by dividing the total market value17 by the total book value18 of the firm at end of the fiscal year (Ting, 2013). There is a statistically significant inverse relationship between CEO turnover and firm performance, which means that CEOs are more likely to be fired when the firm performance is bad than when the firm performance is good (Coughlan & Schmidt, 1985; Defond & Hung, 2004; Jenter & Kanaan, 2008; Warner et al., 1988; Weisbach, 1988). The ratio of Pre-turnover NOI divided by assets is a measure of firm performance and is calculated for the fiscal year prior to the turnover year (Clayton et al., 2005). Clayton et al. (2005) find a significant positive correlation between the share price volatility of a firm and the volatility of the market in which the firm operates. Because the S&P 500 index is used, a control variable for changes in the S&P 500 index volatility is included. The S&P 500 index can be separated in 10 different sectors19. The S&P 500 index volatility is not always a good representation of the volatility of a specific sector. A large volatility change in a S&P 500 sector can be
diversified away in the S&P 500 index by the volatility change of the other S&P 500 sectors. To control this effect, there are nine sector dummies added.
Besides firm and market control variables, the characteristics of a CEO are taken into account. CEO tenure, based on the number of full months being employed, is included as a control variable because as a CEO is employed by the same firm for a longer period of time, investors may become less likely to revise their opinion about the CEO. Often, investors update their prior beliefs after a turnover. The quality of investors updating their beliefs and the tenure of the prior CEO will affect the change in volatility (Clayton et al., 2005). CEO age, in years at the turnover announcement day, is added as control variable because Brickley (2003) concluded that the probability of a dismissal of a CEO is almost 30 percent higher when the CEO is older than 64. A gender dummy, 1 if the CEO is a man and 0 if the CEO is a woman, and a founder dummy, 1 if the CEO is a founder and 0 if the CEO is not a founder, are added to the regression as well.
The criteria of a forced turnover are based on the study of Clayton et al. (2005). The criteria are as followed:
1. It is explicitly stated in the news item that the departing CEO was fired or forced out.
17 Total market value is calculated by multiplying common shares outstanding with the month-end price at the end of the fiscal year.
18 Total book value is total assets minus total liabilities.
19 The 10 S&P 500 sectors are: Consumer Discretionary, Consumer Staples, Energy, Financials, Health Care, Industrials, Information Technology, Materials, Telecommunications Services and Utilities.
2. It is not explicitly stated in the news item that the departing CEO leaves for health reasons or to join another firm.
Next to these criteria, this study will add another criterion to the criteria of Clayton et al. (2005). A CEO may also leave voluntarily due to personal reasons, other than health problems, that have nothing to do with his/her job or the firm.
3. It is not explicitly stated in the news that the departing CEO leaves for personal reasons that are unrelated to the job or firm.
Clayton et al. (2005) also used the following criterion: a CEO turnover is forced when the departing CEO is younger than 60 years old. This criterion does not account for CEOs that retire early. Instead, the following fourth criterion is used:
4. It is explicitly stated in the news item that the departing CEO is younger than 60 years old and the retirement in not planned.
If the turnover satisfies criterion one and/or if the turnover satisfies criteria two, three and four simultaneously, the turnover is classified as forced. Otherwise, the CEO turnover is classified as voluntary. If no news item can be found or if the news item is ambiguous, the turnover is
deleted from the sample.
An event study of a year, defined as 250 trading days before and after the turnover announcement, and an event study of a month, defined as 25 trading days before and after the turnover announcement, will be performed. Because it is often unknown at which moment of the day the announcement is made, the day of the announcement is not included. To estimate the effect of the SOX, each event study can be divided into two periods, namely a 5-year period before the introduction of the SOX and a 5-year period after the introduction of the SOX. As stated, the SOX was introduced in July 2002. This means that it is difficult to include 2002 in one of the two periods. Therefore, the two 5-year periods will be 1997 until 2001 and 2003 until 2007.
The volatility approach, calculating the percentage change of the share price volatility, is used to estimate the effect of a CEO turnover announcement on share price behavior. To apply the volatility approach, for each turnover the standard deviation20 of the share prices must be calculated before and after the turnover announcement day, excluding the
announcement day itself. The percentage difference in volatility before and after the
announcement can be calculated by taking the natural logarithm of the volatility change. The volatility change is calculated by dividing the volatility after the announcement by the
volatility before the announcement. The natural logarithm of the volatility change of each turnover will be regressed on the explanatory and control variables in the statistical program Stata21. The regression will be performed multiple times, for both event studies, by excluding or including some control variable groups to check for robustness.
5. Data
A sample of 150 firms is randomly selected from the S&P 500, using the alphabetic order of the S&P 500 list, by choosing number 1, 5, 10, 15 etc. and number 3, 13, 23, 33 etc. After the selection of these 150 firms, the Execucomp22 annual compensation database is used to determine CEO turnovers. Firms with no turnovers in the sample period and no turnover information in Execucomp are excluded. This leads to a sample of 90 firms and 123 CEO turnovers. The LexisNexis academic23 news database is used to determine the announcement day and to qualify each turnover to forced or voluntary. Figure 1 and figure 2 show the number of forced and voluntary CEO turnovers over the years. The full sample contains 27 forced turnovers (22%) and 96 voluntary turnovers (78%). A similar percentage distribution is found in the period 1997 until 2001 and in the period 2003 until 2007. In the sample period 1997 until 2001 the total number of turnovers, both forced and voluntary, is the highest in 2000. In the sample period 2003 until 2007 the distribution of total turnovers, both forced and voluntary, is approximately the same over the years. There are 59 turnovers, 13 forced and 46 voluntary, before the introduction SOX and there are 64 turnovers, 14 forced and 50
voluntary, after the introduction of the SOX.
21 Stata is a statistical software program in which statistical analysis, regression analysis, data management and simulations can be performed.
22 Execucomp provides data about executive compensation.
23 LexisNexis Academic provides access to newspaper articles, business and legal publications and business information, among others in the areas of finance, business, law and government.
0 2 4 6 8 10 12 14 2003 2004 2005 2006 2007
Figure 2 : Number of CEO turnovers
2003-2007 Forced Voluntary 0 2 4 6 8 10 12 14 16 1997 1998 1999 2000 2001
Figure 1 : Number of CEO turnovers
1997-2001
Data about CEO age, tenure and gender is received from the Execucomp annual compensation database. The regression will be performed without the CEO gender dummy. Only one turnover concerns a woman so including the gender dummy in the regression does not give information about the difference between a man and a woman. To determine if the departing CEO is a founder, the history of the firm is consulted by using the internet website of the firm. Pre-turnover NOI/assets, market-to-book ratio and market capitalization are calculated based on the data of Compustat24 fundamentals annual for North America. Information about the S&P industry sectors can also be found in the Compustat database.
Table 1 shows the information about the characteristics of the control variables, excluding the dummy variables, for the full sample. On average the CEO is around 60 years old at the turnover announcement day and is approximately employed as CEO for 9 years and 4 months. The mean of the market-to-book ratio is 4.11 and the mean of the pre-turnover NOI/assets is 0.15. The market capitalization is on average $ 25881.2 million. Table 2 compares the control variable characteristics of forced turnovers with the control variable characteristics of voluntary turnovers. It can be concluded that a CEO who is forced out is on average younger and employed at the firm for a shorter period of time. When looking at the firm performance control variables, it suggest that firms where forced turnovers occur are generally performing poorly compared to firms where voluntary turnovers occur. The
distribution of the S&P 500 sectors can be found in figure 3 of the appendix. And in the whole sample, 11 (8.9%) CEOs that were founder of the firm left.
24 Compustat provides data about market, financial and statistical information on firms all around the world.
Table 1 : Control variable characterisitcs full sample
Mean SD Minimum Maximum
CEO Age (years) 60.63 7.13 42 83
CEO Tenure (months) 111.95 80.68 4 475
Market-to-Book ratio 4.11 3.50 -0.19 25.26
Market Capitalization (mln. $) 25,881.20 57,500.80 463.90 460,767.90
Pre-Turnover NOI/Assets 0.15 0.09 -0.17 0.43
Table 2 : Control variable characterisitcs voluntary and forced sample
Voluntary Forced
Mean SD Mean SD
CEO Age (years) 61.75 6.17 56.63 8.85
CEO Tenure (months) 120.05 79.71 83.15 78.88
Market-to-Book ratio 4.27 3.66 3.56 2.85
Market Capitalization (mln. $) 27,808.84 63,142.11 19,027.37 29,728.70
The CRSP25 daily stock database is used to find historical firm and S&P 500 index share prices to calculate the volatility. In table 3 the ratio of the volatility change, the standard deviation before a turnover announcement divided by the standard deviation after a turnover announcement, are represented for both event studies. In the one-year event study, there is an increase in the standard deviation after a turnover announcement for all sample groups, except for pre-SOX and forced sample. However, in the one-month event study the standard
deviation decreases after a turnover announcement, except for the forced and the pre-SOX and forced sample. This implies that volatility decreases in a short time period and volatility increases in a longer time period volatility increases after a CEO turnover announcement. The standard deviation of a forced turnover increases in both event studies. However, this is not the case when the forced turnover occurs in the pre-SOX period for the one-year event study or in the SOX period for the one-month event study. For both the pre-SOX and post-SOX periods, the standard deviation increases in the one-year event study and decreases in the one-month event study. Figure 4 shows that there are no large differences between percentage volatility changes in the post- and pre-SOX period for the one-year period. Table 4.1 and table 4.2 of the appendix give a broad overview of the mean volatility pre- and post-turnover for the different sample groups in both event studies.
The difference between the mean standard deviation before and after the turnover announcement is presented in table 5.1 and table 5.2. In the one-year event study, the volatility change is the largest for the full sample (0.580%), the voluntary sample (0.729%) and the pre-SOX sample (0.854%). In the event study of a month the volatility change of the voluntary sample (-0.147%), post-SOX sample (-0.152%) and the pre-SOX and forced sample (0.261%) are the largest. In the one-year event study, the volatility change for the full sample, the
25 CRSP provides data about historical share prices, returns and volume on the NYSE, AMEX, and NASDAQ markets.
Table 3 : Ratio of volatility change
One month One year
Full sample 0.935 1.136
Forced turnover 1.044 1.012
Voluntary turnover 0.911 1.171
Pre-SOX 0.976 1.173
Pre-SOX and Forced 1.153 0.988
post-SOX 0.868 1.090
Post-SOX and Forced 0.847 1.045
voluntary sample and the S&P 500 index are significant. The volatility change in the pre-SOX period is only significant using a paired t-test. The volatility change in the post-SOX period is only significant using a Wilcoxon signed-rank test. In the one-month event study, no volatility change is significant, since they are too small to be significant.
6. Results
For a first impression of the results, the outcomes of a two-sample t-test for differences in volatility changes are presented in table 6.1 and 6.2. It can be concluded that the volatility of a voluntary turnover increases significantly 0.295% more for the one-year period and increases insignificantly 0.051% less for the one-month period compared to a forced turnover after the CEO turnover announcement. This implies that the first hypothesis is not rejected for the one-year event study, the volatility change for a voluntary turnover is significantly larger than the volatility change for a forced turnover after a CEO turnover announcement. The first
hypothesis is not rejected for the one-month event study either because the difference is insignificant. It can also be concluded that on average the volatility change of a forced
Table 5.1 : Difference in volatilty one-year period
Difference between means Paired t-test Wilcoxon Z value
Full sample 0.580 2.297** 2.350**
Forced turnover 0.050 0.096 -0.456
Voluntary turnover 0.729 2.529** 2.902***
Pre-SOX 0.854 1.936* 1.540
Pre-SOX and forced -0.060 -0.068 -0.664
Post-SOX 0.328 1.237 1.759*
Post-SOX and forced 0.152 0.253 0.282
S&P 500 index 6.532 2.275** 2.259**
All differences are in percentage * means a significance level of 10% ** means a significance level of 5% *** means a significance level of 1%
Table 5.2 : Difference in volatilty one-month period
Difference between means Paired t-test Wilcoxon Z value
Full sample -0.102 -0.968 -0.237
Forced turnover 0.056 0.342 0.889
Voluntary turnover -0.147 -1.152 -0.771
Pre-SOX -0.048 -0.262 0.702
Pre-SOX and forced 0.261 1.213 1.153
Post-SOX -0.152 -1.343 -1.284
Post-SOX and forced -0.135 -0.563 0.282
S&P 500 index -1.251 -1.319 -1.401
turnover moves in the opposite direction compared to the volatility change of a voluntary turnover for both event studies. This is evident in the positive and negative signs in respectively the one-year event study and the one-month event study.
This study finds the volatility change is larger in the pre-SOX period compared to the post-SOX period, namely 0.022% larger in the one-year event study and 0.098% larger in the one-month event study.But this small difference is insignificant. When looking at only the forced sample, the one-year volatility for the pre-SOX period decreases 0.122% more compared to the post-SOX period after a turnover announcement. In the one-month event study, the volatility in the pre-SOX period increases 0.019% more compared to the pre-SOX period after a turnover announcement. The volatility change, whether positive ornegative, is in both event studies smaller in the post-SOX period compared to the pre-SOX period for forced turnover only. As expected, there is a difference in volatility change before and after the introduction of the SOX in the whole sample and in the forced sample. However, these differences are not significant. On this ground, the second hypothesis cannot be rejected using a two-sample t-test for differences in volatility changes. No conclusion can be made yet, from table 6.1 and 6.2, about hypothesis three because a one-sided t-test is needed.
In table 7.1 and table 7.2 the results of the regression with only the variables of
interest, the forced, SOX and Forced*SOX coefficients, are presented. The first interpretation of these results suggests that there exists a significant negative effect of a forced turnover, compared to a voluntary turnover, on the volatility change for the one-year event study. There is an insignificant negative effect of the post-SOX period, compared to the pre-SOX period, on the volatility change for the one-month event study. No other positive or negative effects appear from this regression.
Table 6.1 : Difference in volatility change one-year period
(1 vs. 2) Mean volatility change(1) Mean volatility change(2) Difference Std. Err. T-test N
Voluntary vs. Forced 0.146 -0.148 0.295* 0.117 2.512 27/96
Pre-SOX vs. Post-SOX (whole sample) 0.093 0.071 0.022 0.100 0.224 59/64
Pre-SOX vs. Post-SOX (only forced) -0.212 -0.090 -0.122 0.234 -0.520 13/14 All mean volatility changes are in percentage
* means a significance level of 5%
Table 6.2 : Difference in volatility change one-month period
(1 vs. 2) Mean volatility change(1) Mean volatility change(2) Difference Std. Err. T-test N
Voluntary vs. Forced -0.049 0.003 -0.051 0.142 -0.363 27/96
Pre-SOX vs. Post-SOX (full sample) 0.013 -0.084 0.098 0.117 0.834 59/64
Pre-SOX vs. Post-SOX (only forced) 0.012 -0.006 0.019 0.279 0.068 13/14
Subsequently, the regression is performed multiple times, for both event studies, by excluding or including some control variable groups to check for robustness. The results are shown in table 8.1 and 8.2. Regression nine has the highest R2 value in the both event studies. Therefore, this regression explains the most about the variability in volatility changes after a CEO turnover announcement26. Hence, regression nine is used to test the hypotheses.
In all regressions in the one-year event study the coefficient of the forced dummy is significantly negative. The significance level of the forced dummy coefficient is even higher when the sector dummies are excluded from the regression, namely in regression one, two, four and five. The t-value of the forced dummy coefficient in regression nine is -1.76, which is smaller than 1.657, H0 is not rejected. Therefore, hypothesis one cannot be accepted in the one-year period27. Instead this study finds that the coefficient is significantly smaller than
26 R2 tells how much of the volatility change is explained by the variables in the model. If R2 is zero, nothing is explained and if R2 is 1, everything is explained.
27 T-value = -0.364 / 0.207 = -1.76
Table 7.1 : Regression without control variables one-year period
Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Forced -0.295 ** -0.295** -0.358** -0.391** (0.126) (0.127) (0.163) (0.175) SOX -0.022 -0.023 0.023 -0.063 (0.100) (0.098) (0.103) (0.108) Forced*SOX -0.193 0.122 -0.205 0.185 (0.174) (0.227) (0.181) (0.252) Constant 0.146*** 0.093 0.103* 0.158** 0.146*** 0.093 0.179 (0.053) (0.075) (0.052) (0.077) (0.053) (0.075) (0.082) N 123 123 123 123 123 123 123 R2 0.050 0.000 0.013 0.050 0.052 0.013 0.105 F-test 5.460** 0.055 1.230 2.700 3.040 0.650 2.09
* means a significance level of 10% ** means a significance level of 5% *** means a significance level of 1%
Robust standard errors are given in parentheses.
Table 7. 2 : Regression without control variables one-month period
Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Ln(SDt+1/SDt-1) Forced 0.051 0.051 0.061 -0.001 (0.150) (0.151) (0.178) (0.193) SOX -0.098 -0.098 -0.119 -0.120 (0.117) (0.118) (0.120) (0.131) Forced*SOX 0.035 -0.019 0.100 0.101 (0.220) (0.269) (0.229) (0.300) Constant -0.049 0.013 -0.041 0.002 -0.049 0.013 0.014 (0.065) (0.084) (0.060) (0.092) (0.065) (0.084) (0.098) N 123 123 123 123 123 123 123 R2 0.001 0.006 0.000 0.050 0.001 0.008 0.008 F-test 0.120 0.700 0.030 2.700 0.070 0.510 0.340
* means a significance level of 10% ** means a significance level of 5% *** means a significance level of 1%
zero, at a significance level of 5% 28. This means that the volatility change of a forced
turnover is significantly smaller than volatility change of a voluntary turnover in the one-year event study. This conclusion is opposite to the findings of Clayton et al. (2005), who find an increase of the volatility change for forced turnovers compared to voluntary turnovers. And this conclusion is in line with the conclusion of Ting (2013), who finds a decrease of the volatility change for forced turnovers compared to voluntary turnovers. However, in the one-month event study another conclusion is drawn. In the nine regressions for the one-one-month period, there is no unambiguous positive or negative direction of the forced dummy
coefficient. All forced dummy coefficients are insignificant. For regression nine H0 cannot be rejected, with a t-value of 0.0829. This means that in the shorter run there is no significant difference between the volatility change of a forced and a voluntary turnover.
In all regressions of the one-year event study the SOX dummy is negative. This means that the volatility change is smaller in the post-SOX period than in the pre-SOX period after a turnover announcement. This negative effect is the largest for regression one and four.
However, these nine different SOX dummy coefficients are not significant. H0 is not rejected in all regressions, therefore hypothesis two cannot be accepted for the one-year period. This unambiguous negative direction is larger for the one-month period, especially in regression two. However, in all regressions, this coefficient is still not significant. In conclusion, in line with the expectations there is an effect of the SOX on the volatility change but this negative effect is not significantly different from zero in both periods.
When the SOX dummy interacts with the forced dummy, a positive effect is found for the one year period. Especially in the eighth and ninth regression. However, these coefficients are not significant. Therefore, H0 is not rejected and hypothesis three cannot be accepted for the one-year period. However, it can be suggested that this positive effect is economically significant30 because of the magnitude and unambiguous sign of the coefficient in all
regressions. For the one-month period, the regressions do not give an unambiguous negative or positive direction. This has nothing to do with whether the control variable groups are included or excluded. In this case, H0 is not rejected either.
28 H
0: β1 = 0 H1: β1 < 0 Critical value: t < -1.657
Ta b le 8 .1 : F u ll re g re ss io n o n e -y e a r p e ri o d L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) F o rc e d -0 .4 1 3 * * -0 .3 8 3 * * -0 .4 0 1 * -0 .3 8 7 * * -0 .4 0 8 * * -0 .3 6 5 * -0 .3 9 6 * -0 .3 6 8 * -0 .3 6 4 * (0 .1 7 7 ) (0 .1 8 0 ) (0 .1 8 0 ) (0 .1 7 8 ) (0 .1 8 6 ) (0 .1 9 8 ) (0 .2 1 8 ) (0 .2 0 4 ) (0 .2 0 7 ) S O X -0 .0 6 4 -0 .0 5 6 -0 .0 6 2 -0 .0 6 5 -0 .0 6 1 -0 .0 4 1 -0 .0 6 0 -0 .0 4 0 -0 .0 3 9 (0 .1 1 2 ) (0 .1 1 0 ) (0 .1 0 5 ) (0 .1 0 8 ) (0 .1 1 8 ) (0 .1 0 9 ) (0 .1 1 1 ) (0 .1 1 4 ) (0 .1 1 4 ) F o rc e d * S O X 0 .2 0 7 0 .1 8 6 0 .2 5 2 0 .1 8 5 0 .2 2 2 0 .2 7 0 0 .2 5 5 0 .2 8 1 0 .2 7 9 (0 .2 6 3 ) (0 .2 5 7 ) (0 .2 6 9 ) (0 .2 5 3 ) (0 .2 7 2 ) (0 .2 7 4 ) (0 .2 7 6 ) (0 .2 8 0 ) (0 .2 8 2 ) CE O a g e -0 .0 0 2 -0 .0 0 3 0 .0 0 4 0 .0 0 3 0 .0 0 4 (0 .0 0 7 ) (0 .0 0 7 ) (0 .0 0 8 ) (0 .0 0 8 ) (0 .0 0 8 ) CE O fo u n d e r 0 .1 8 2 0 .2 3 5 0 .0 9 7 0 .1 2 5 0 .1 2 7 (0 .1 9 9 ) (0 .2 0 5 ) (0 .2 0 4 ) (0 .2 0 7 ) (0 .2 1 1 ) CE O t e n u re -0 .0 0 0 -0 .0 0 0 0 .0 0 0 0 .0 0 0 0 .0 0 0 (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) L n (S & P v o la ti lit y t + 1 /S & P v o la ti lit y t -1 ) -0 .0 3 6 -0 .0 4 9 -0 .0 4 7 -0 .0 4 2 -0 .0 5 1 (0 .1 0 9 ) (0 .1 1 9 ) (0 .1 2 5 ) (0 .1 2 5 ) (0 .1 2 8 ) P re -t u rn o v e r N O I/ a ss e ts 0 .4 8 9 0 .6 5 9 0 .4 2 4 0 .4 7 1 0 .4 9 0 (0 .5 4 6 ) (0 .5 7 0 ) (0 .5 3 1 ) (0 .5 1 8 ) (0 .5 2 9 ) M a rk e t-t o -b o o k ra ti o -0 .0 0 6 -0 .0 1 1 -0 .0 0 5 -0 .0 0 6 -0 .0 0 5 (0 .0 1 0 ) (0 .0 1 1 ) (0 .0 1 2 ) (0 .0 1 2 ) (0 .0 1 2 ) ln (M rk t. Ca p .) -0 .0 1 0 -0 .0 1 0 0 .0 0 7 0 .0 0 8 0 .0 0 8 (0 .0 4 0 ) (0 .0 4 1 ) (0 .0 4 0 ) (0 .0 4 2 ) (0 .0 4 2 ) In d u st ry d u m m ie s No No Y e s No No Y e s Y e s Y e s Y e s c o n st a n t 0 .3 2 2 0 .2 1 6 0 .1 5 8 0 .1 8 3 * * 0 .3 7 0 -0 .1 3 7 0 .0 5 2 -0 .2 1 1 -0 .2 3 3 (0 .4 0 0 ) (0 .4 0 5 ) (0 .2 0 5 ) (0 .0 8 3 ) (0 .6 6 1 ) (0 .5 4 2 ) (0 .4 5 0 ) (0 .7 8 0 ) (0 .8 0 8 ) N 123 123 123 123 123 123 123 123 123 R2 0 .0 6 5 0 .0 6 0 0 .1 4 7 0 .0 5 6 0 .0 7 6 0 .1 5 7 0 .1 5 2 0 .1 6 0 0 .1 6 2 F -t e st 1 .5 8 1 .2 2 3 .2 4 * * * 1 .5 5 1 .1 6 2 .6 6 * * * 2 .5 6 * * * 2 .4 5 * * * 2 .3 6 * * * * m e a n s a s ig n ifi c a n c e l e v e l o f 1 0 % * * m e a n s a s ig n ifi c a n c e l e v e l o f 5 % * * * m e a n s a s ig n ifi c a n c e l e v e l o f 1 % Ro b u st s ta n d a rd e rro rs a re g iv e n i n p a re n th e se s.
Ta b le 8 .2 : F u ll re g re ss io n o n e -m o n th p e ri o d L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) L n (S D t+ 1 /S D t-1 ) F o rc e d 0 .0 6 3 -0 .0 8 1 0 .0 6 8 -0 .0 2 5 -0 .0 2 2 0 .0 5 3 -0 .0 1 0 0 .0 4 1 0 .0 1 9 (0 .2 0 2 ) (0 .2 0 0 ) (0 .2 0 1 ) (0 .1 9 4 ) (0 .2 1 3 ) (0 .2 1 8 ) (0 .2 1 4 ) (0 .2 3 2 ) (0 .2 3 3 ) S O X -0 .0 9 0 -0 .1 9 5 -0 .1 1 7 -0 .1 2 3 -0 .1 5 5 -0 .1 0 7 -0 .1 7 5 -0 .1 5 2 -0 .1 5 2 (0 .1 4 3 ) (0 .1 3 2 ) (0 .1 3 8 ) (0 .1 2 9 ) (0 .1 4 7 ) (0 .1 5 3 ) (0 .1 5 2 ) (0 .1 7 2 ) (0 .1 7 1 ) F o rc e d * S O X 0 .0 4 9 0 .1 8 4 -0 .0 4 3 0 .1 3 4 0 .1 5 2 -0 .0 3 2 0 .0 8 1 -0 .0 1 0 0 .0 3 8 (0 .3 0 0 ) (0 .2 9 7 ) (0 .3 1 5 ) (0 .2 9 1 ) (0 .2 8 9 ) (0 .3 0 7 ) (0 .3 0 3 ) (0 .3 1 7 ) (0 .3 0 8 ) CE O a g e 0 .0 0 2 0 .0 0 3 -0 .0 0 5 -0 .0 0 6 -0 .0 0 5 (0 .0 1 0 ) (0 .0 0 9 ) (0 .0 1 1 ) (0 .0 1 1 ) (0 .1 1 ) CE O fo u n d e r -0 .2 9 0 -0 .2 3 0 -0 .1 0 0 -0 .2 0 8 -0 .1 3 0 (0 .2 0 1 ) (0 .2 1 7 ) (0 .2 1 5 ) (0 .2 1 6 ) (0 .2 2 2 ) CE O t e n u re 0 .0 0 1 0 .0 0 1 0 .0 0 1 0 .0 0 1 0 .0 0 1 (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) (0 .0 0 1 ) L n (S & P v o la ti lit y t + 1 /S & P v o la ti lit y t -1 ) 0 .2 8 7 * * * 0 .2 3 3 0 .2 3 9 * 0 .2 3 4 * 0 .2 1 7 (0 .1 2 8 ) (0 .1 3 0 ) (0 .1 3 6 ) (0 .1 3 4 ) (0 .1 3 4 ) P re -t u rn o v e r N O I/ a ss e ts -0 .8 4 8 -0 .7 6 4 -0 .3 7 7 -0 .6 8 9 -0 .4 5 3 (0 .7 3 2 ) (0 .7 3 0 ) (0 .7 8 2 ) (0 .7 8 1 ) (0 .8 2 2 ) M a rk e t-t o -b o o k ra ti o -0 .0 1 7 -0 .0 0 8 -0 .0 0 8 -0 .0 0 6 -0 .0 0 5 (0 .0 1 7 ) (0 .0 1 8 ) (0 .0 1 7 ) (0 .0 1 8 ) (0 .0 1 8 ) ln (M rk t. Ca p .) 0 .0 7 9 * 0 .0 7 3 0 .0 6 1 0 .0 6 8 0 .0 6 3 (0 .0 4 3 ) (0 .0 4 3 ) (0 .0 4 5 ) (0 .0 4 7 ) (0 .0 4 6 ) In d u st ry d u m m ie s No No Y e s No No Y e s Y e s Y e s Y e s c o n st a n t -0 .2 0 7 -0 .4 5 7 0 .2 2 0 0 .0 2 8 -0 .7 2 1 0 .4 5 7 -0 .2 8 0 .0 5 1 -0 .0 6 7 (0 .5 4 3 ) (0 .4 1 4 ) (0 .2 0 1 ) (0 .9 9 0 ) (0 .7 2 5 ) (0 .6 6 6 ) (0 .4 3 4 ) (0 .8 7 0 ) (0 .8 5 9 ) N 123 123 123 123 123 123 123 123 123 R2 0 .0 2 7 0 .0 5 2 0 .0 8 7 0 .0 5 4 0 .0 9 8 0 .1 2 7 0 .1 3 7 0 .1 1 9 0 .1 4 3 F -t e st 1 .0 4 1 .3 6 1 .7 8 * 1 .6 9 1 .8 4 * 1 .5 6 * 1 .7 5 * * 3 .5 3 * * * * 1 .6 8 * * m e a n s a s ig n ifi c a n c e l e v e l o f 1 0 % * * m e a n s a s ig n ifi c a n c e l e v e l o f 5 % * * * m e a n s a s ig n ifi c a n c e l e v e l o f 1 % Ro b u st s ta n d a rd e rro rs a re g iv e n i n p a re n th e se s.
7. Discussion
What stands out is that all regressions, in both event studies, have a small R2. However, the object of this study is not to explain all variability in the volatility changes but to find the effects of turnovers on the volatility change. Therefore, it is not relevant that the regression does not explain most of the volatility change. The low R2 is comparable to the studies of Clayton et al. (2005) and Ting (2013), even though they both used larger samples of respectively 872 and 397. This low R2 results in small F-test values for the control variable groups in both event studies, as shown in table 9.1 and table 9.2 of the appendix. Only the S&P sector dummies in the one-year event study are simultaneously significantly different from zero.
In the one-month time period, no ambiguous direction for the forced dummy
coefficient is found. This implies that there is no direct short term effect of a forced turnover on the volatility change. However, in the one-year time period the coefficient is significantly smaller than zero. According to Huson et al. (2004), a CEO who is fired due to poor
performance can be seen as a scapegoat, because poor performance is a result of bad luck31. A forced CEO turnover will not increase the managerial quality of the CEO, since they employ all at the same optimal level, but the expected change in firm performance following forced turnover is positive (Huson et al., 2004). This indicates that in the long run the
uncertainty about future firm performance decreases, this leads to the significant negative forced dummy coefficient that is found in the one-year event study.
Generally, the SOX has a negative insignificant effect on the volatility change in both event studies. It can be suggested that the reaction of investors to a CEO turnover is
diminished because in both periods the volatility change decreased after the introduction of the SOX. This implies that the investors have more confidence in the internal control mechanisms of the firm in combination with the SOX. The decrease in volatility change is possibly a result of investors believing that the SOX will better align the principals’ objectives with the objectives of the agent (Hochberg et al., 2009). If the statistical evidence for the positive view about the SOX, that is found by Hochberg et al. (2009), applies to shareholders of S&P 500 firms, the negative insignificant effect can possibly be explained by the better alignment of the interests. There will probably be a decrease in the uncertainty about the alignment of the interest of the shareholders and the new CEO and this will probably reduce uncertainty about future shareholders’ value. Consequently, reducing the volatility change
after the introduction of the SOX. However, the existence of this negative volatility effect has to be taken in consideration because no statistical evidence is found in this study.
In the post-SOX period, earnings management remains significantly positively related to the likelihood of a forced CEO turnover and unrelated to a voluntary CEO turnover
(Hazarika et al., 2012). According to the findings of Li et al. (2008), investors expect that the SOX will limit earnings management and improve the quality of financial statement
information. In the one-year event study a positive effect on the volatility change is found when the turnover is forced and occurs in post-SOX period. However, the SOX has no effect on the volatility of forced turnovers within the one-month event study. This implies that in the short term there is no difference in volatility change after a forced turnover between the pre-SOX and post-pre-SOX period. However, in the longer term uncertainty may increase on the efficiency of the SOX to stimulate CEOs not to engage in earnings management and to improve the quality of the financial statements of the firm. This uncertainty may increase the volatility in the one-year period because shareholders doubt about the quality of the financial performances that are reported after the turnover announcement. On the other hand, the objectives of the SOX suggest that the board, and therefore the CEO as leader of the company must be subject to rules and control mechanisms to prevent fraud and to protect shareholders’ value. The SOX could cause to more awareness of shareholders to financial information about the firm or to information about the quality or performance of the CEO and the board after a forced turnover announcement. This will lead to a positive effect on the volatility change after a forced turnover in the post-SOX period.
8. Conclusion
This study follows Clayton et al. (2005) and Ting (2013) to analyze the announcement effect of a CEO turnover based on a sample of 150 random selected S&P 500 firms. Rather than looking at the abnormal cumulative stock returns, volatility changes are calculated before and after the day of the turnover announcement to analyze the announcement effect. This
approach is called the volatility approach. In total, 123 turnovers are examined Hereof, 27 turnovers (22%) are classified as forced and 96 turnovers (78%) are classified as voluntary.
In the one-year period, a robust significant negative effect on the volatility change is found after a forced turnover compared to a voluntary turnover. This result is in line with the findings of Ting (2013), who finds a significant negative forced CEO turnover effect on the volatility change compared to voluntary turnovers. An explanation for this result is that forced
CEO turnover does not increase the managerial quality of the CEO but the expected change in firm performance following forced turnover is positive (Huson et al., 2004). In the one-month period, no significant difference between the volatility change after a forced or a voluntary turnover announcement is found. The findings indicate that a forced turnover has no direct effect, but does have a long-term effect on the volatility by decreasing the uncertainty about future firm performance.
The purpose of this study is to investigate a possible difference in share price volatility change, after a CEO turnover announcement, before and after the introduction of the SOX. In addition to this, this study investigates whether this effect is greater when the announcement concerns a forced turnover in respect to a voluntary turnover. This study contributes in the understanding of the impact of the SOX on the behavior of shareholders. In the sample there are 59 turnovers in the pre-SOX period and 64 turnovers in the post-SOX period. There is no significant difference found in share price volatility before and after the introduction of the SOX.
There is found an insignificant negative SOX effect on the volatility change in both event study periods. A possible explanation of this negative insignificant effect is that investors believe that the SOX will better align the shareholders’ objectives with the objectives of the new CEO (Hochberg et al., 2009). This will probably reduce uncertainty about future shareholders’ value and reducing the volatility change after the introduction of the SOX. However, the existence of this negative volatility effect has to be taken in
consideration because no statistical evidence is found.
A statistical insignificant but economically significant positive SOX effect is found for forced CEO turnover announcements in the post-SOX period for the one-year period. No unambiguous significant effect of forced turnovers in the post-SOX period is found in the one-month period. This long term positive effect may be a result of the increased uncertainty about the efficiency of the SOX to stimulate CEOs not to engage in earnings management and to improve the quality of the financial statements of the firm. This possibly increases the volatility in the one-year period because shareholders doubt about the quality of the financial performances that are reported after the turnover announcement. Another explanation could be that the SOX could cause more awareness of shareholders to financial information about the firm or to information about the quality or performance of the CEO and the board after a
9. References
Adams, J. C., & Mansi, S. A. (2009). CEO turnover and bondholder wealth. Journal of Banking & Finance, 33(3), 522-533.
Bertrand, M., & Schoar, A. (2003). Managing with style: The effect of managers on firm policies. The Quarterly Journal of Economics, 118(4), 1169–1208.
Bonnier, K. A., & Bruner, R. F. (1989). An analysis of stock price reaction to management change in distressed firms. Journal of Accounting and Economics, 11(1), 95-106. Brickley, J. A. (2003). Empirical research on CEO turnover and firm performance: a
discussion. Journal of Accounting and Economics, 36, 227-233.
Chemmanur, T. J., & Fedaseyeu, V. (2012). A theory of corporate boards and forced CEO turnover. Retrieved from http://ssrn.com/abstract=1647645
Clayton, M. C., Hartzell, J. C., & Rosenberg, J. (2005). The impact of CEO turnover on equity volatility. The Journal of Business, 78(5), 1779-1808.
Coughlan, A.T., & Schmidt, R. M. (1985). Executive compensation, management turnover, and firm performance; an empirical investigation. Journal of Accounting and
Economics, 7, 43-66.
Dah, M. A., Frye, M. B., & Hurst, M. (2014). Board changes and CEO turnover: The
unanticipated effects of the Sarbanes–Oxley Act. Journal of Banking and Finance, 41, 97-108.
Davidson, W. N., Wang, H., & Wang, X. (2010). The Sarbanes-Oxley Act and CEO
tenure, turnover, and risk aversion. The Quarterly Review of Economics and Finance, 50(3), 367-376.
Dedman, E., & Lin, S.W. J. (2002). Shareholder Wealth Effects of CEO Departures: Evidence from the UK. Journal of corporate finance, 8(1), 81-104.
Defond, M. L., & Hung, M. (2004). Investor protection and corporate governance: Evidence from worldwide CEO turnover. Journal of Accounting Research, 42(2), 269-312. Filbeck, G., Gorman, R., & Zhao, X. (2011). SOX and the regulated firm. Journal of
Accounting and Public Policy, 30(6), 526-550.
Hazarika, S., Karpoff, J. M., & Nahata, R. (2012). Internal corporate governance, CEO turnover, and earnings management. Journal of Financial Economics, 104(1), 44-69. Hochberg, Y. V., Sapienza, P., & Vissing‐Jørgensen, A. (2009). A Lobbying Approach to
Evaluating the Sarbanes‐Oxley Act of 2002. Journal of Accounting Research, 47(2), 519-583.
Huson, M. R., Malatesta, P. H., & Parrino, R. (2004). Managerial succession and performance. Journal of Financial Economics, 74(2), 237-275.
Huson, M. R., Parrino, R., & Starks, L. (2001). Internal monitoring mechanisms and CEO turnover: A long-term perspective. Journal of Finance, 56, 2265–2297.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360. Jenter, D., & Kanaan, F. (2008). CEO turnover and relative performance evaluation.
Retrieved from http://ssrn.com/abstract=885531
Kaplan, S. N., & Minton, B. A. (2012). How has CEO turnover changed? International Review of Finance, 12(1), 57-87.
Khanna, N., & Poulsen, A. B. (1995). Managers of financially distressed firms: villains or scapegoats? The Journal of Finance, 50(3), 919-940.
Kind, A., & Schläpfer, Y. (2011). Are Forced CEO Turnovers Good or Bad News? Retrieved from http://ssrn.com/abstract=1679632
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (2000). Investor protection and corporate governance. Journal of financial economics, 58(1), 3-27.
Li, H., Pincus, M., & Rego, S. (2008). Market reaction to events surrounding the Sarbanes– Oxley Act of 2002 and earnings management. Journal of Law and Economics, 51(1), 111–134.
Linck, J. S., Netter, J. M., & Yang, T. (2009). The effects and unintended consequences of the Sarbanes-Oxley Act on the supply and demand for directors. Review of Financial Studies, 22(8), 3287-3328.
Ting, H. I. (2013). CEO turnover and shareholder wealth: evidence from CEO power in Taiwan. Journal of Business Research, 66(12), 2466-2472.
Warner, J. B., Watts, R. L., & Wruck, K. H. (1988). Stock prices and top management changes. Journal of Financial Economics, 20, 461-192.
Weisbach, M. (1988). Outside directors and CEO turnover. Journal of Financial Economics, 20, 431–460.
Zhang, I. X. (2007). Economic consequences of the Sarbanes–Oxley Act of 2002. Journal of Accounting and Economics, 44(1), 74-115.
10. Appendix 0 4 8 12 16 20 24 28
Figure 3: S&P sectors
Table 4.1 : Summary statistics one-year period
Pre-turnover SD of firm returns
Mean SD Minimum Maximum N
Full sample 4.265 2.671 0.468 12.760 123
Forced turnover 4.247 2.858 0.721 12.760 27
Voluntary turnover 4.270 2.632 0.468 12.645 96
Pre-SOX 4.945 2.919 0.468 12.760 59
Pre-SOX and Forced 5.206 3.010 1.826 12.760 13
post-SOX 3.637 2.267 0.721 11.288 64
Post-SOX and Forced 3.356 2.490 0.721 10.294 14
S&P 500 index 61.307 23.261 14.411 122,885 123
Post-turnover SD of firm returns
Full sample 4.845 3.612 0.595 21.403 123
Forced turnover 4.297 4.378 0.595 19.596 27
Voluntary turnover 4.999 3.376 0.939 21.403 96
Pre-SOX 5.799 4.398 1.065 21.403 59
Pre-SOX and forced 5.146 5.263 1.244 19.596 13
Post-SOX 3.965 2.412 0.595 11.918 64
Post-SOX and forced 3.508 3.372 0.595 11.918 14
S&P 500 index 67.840 25.967 22.342 168.893 123
Table 4.2 : Summary statistics one-month period
Pre-turnover SD of firm returns
Mean SD Minimum Maximum N
Full sample 1.564 1.385 0.156 8.637 123
Forced turnover 1.280 1.099 0.233 4.965 27
Voluntary turnover 1.644 1.450 0.156 8.637 96
Pre-SOX 2.015 1.85 0.156 8.637 59
Pre-SOX and Forced 1.709 1.213 0.373 4.965 13
post-SOX 1.149 0.856 0.188 4.002 64
Post-SOX and Forced 0.825 0.837 0.233 2.714 14
S&P 500 index 22.367 10.566 7.878 52.761 123
Post-turnover SD of firm returns
Full sample 1.462 1.182 0.217 6.802 123
Forced turnover 1.336 1.352 0.217 6.802 27
Voluntary turnover 1.497 1.135 0.234 5.402 96
Pre-SOX 1.967 1.387 0.297 6.802 59
Pre-SOX and forced 1.970 1.708 0.297 6.802 13
Post-SOX 0.997 0.690 0.217 4.408 64
Post-SOX and forced 0.748 0.435 0.217 1.587 14
S&P 500 index 21.116 9.326 6.204 59.496 123
All standard deviations are in percentage
Table 9.1 : Control variable groups
one-year period F-test Ceo characteristics 0.40 Firm characteristics 0.22 S&P 500 volatility 0.16 S&P 500 sectors 2.43*
* means a significance level of 5%
Table 9.2: Control variable groups
one-month period F-test Ceo characteristics 0.19 Firm characteristics 0.68 S&P 500 volatility 2.62 S&P 500 sectors 1.25
