The effect of CEO seniority on diversifying acquisition returns

(1)

1

MSc Business Economics, Finance track

Master Thesis

The effect of CEO seniority on diversifying acquisition returns

Abstract

This research shows how CEO seniority affects diversifying acquisition’s announcement day returns in the period 1995-2014. The research shows that CEOs have significant compensation incentives in pursuing acquisitions. The compensation incentive is even stronger for diversifying acquisitions. The compensation incentive creates possible agency costs. The results show that diversifying acquisition returns increase with CEO age. The positive effects appears to be strongest for middle aged CEOs. I argue that younger CEOs are driven by the compensation incentive and not by returns in diversifying acquisitions. The younger CEOs outperform older CEOs in focused acquisitions. This research shows the importance of CEO seniority regarding acquisition decisions and possible corporate governance problems.

Ward van den Boorn 6062768

July 2015

(2)

2

The separation of ownership and control began when small firms started growing into large corporations that could not be controlled by just the original owner. In 1992-2006 executives of firms own on average 2.8% of the outstanding shares, while they do manage the firm (Kim and Lu, 2011). The separation of ownership and control is in many ways beneficial. However, it has disadvantages as well. It is widely known that the separation of ownership and control can lead to several corporate governance issues (Fama & Jensen, 1983). The decisions that lead to these issues are taken by CEOs for their own benefit and might cause a loss of value for shareholders. That corporate governance issues lead to costs for firms is shown in a research by McKinsey in 20021. This research shows that 80% of the investors state they are willing to pay a premium for a good governed firm.

One of the corporate governance issues is empire building. This means a CEO is focused on expanding the corporation to gain power, prestige and compensation. For shareholders, this expansion drift is not always beneficial. Shareholders try to diminish the abuse of control by CEOs through good corporate governance regulation. Masulis et al (2007) show that firms with good corporate governance suffer less from managerial entrenchment that leads to empire building. However, it cannot be prevented entirely that CEOs might use their control for personal benefit. It is interesting to know what factors influence empire building and when the risk for this corporate governance problem is higher.

In recent research the focus on determinants of acquisition returns is shifting from firm specific to governance specific. The fact that investors are willing to pay for good governance makes it an interesting field of research. This research will focus on CEO seniority and the effect on diversifying acquisition returns. I use the assumption that diversifying acquisitions harm shareholders and firm value, as shown in earlier research2. This paper tests how the market reacts on the announcement of diversifying acquisitions and the relation with CEO age. This research will test if younger CEOs generate lower announcement returns in diversifying acquisitions than older CEOs. As Yim (2013) argues, there are incentives for younger CEOs to pursue empire building to increase their compensation. Earlier research also shows a theory called the horizon problem3. CEOs nearing retirement have a shorter horizon than the shareholders of the firm. This shorter horizon can lead to decisions for the CEO’s personal benefit. This leads to agency costs. These personal benefits are mostly the avoidance of risks.

1_{See Larcker & Tayan (2011)} 2

See Berger and Ofek (1995), Hoechle et al (2012), Lamont and Polk (2002), Malmendier and Tate (2008) and Morck et al (1990)

(4)

4 This research will contribute to the already existing literature in several ways. Earlier research covers the effect of multiple CEO characteristics on acquisition returns4. However, the effect of age is relatively under researched. Yim (2013) covers an aspect in this field but her main focus is on the relation between acquisition propensity and CEO age. She does not find a significant effect on acquisition returns and makes no separation in the kind of acquisition. By including this separation between diversifying and focused acquisitions I show significant results. By testing whether CEOs have compensation driven incentives in pursuing acquisitions this paper can provide information about optimizing CEO incentives. Overall this paper can offer valuable insights into determinants of M&A success. The result of this research could be beneficial for governance regulation in corporations. It can be a guideline in deciding how much regulation and guidance the CEO needs in pursuing his acquisitions.

The results show that acquisition activity has a significant effect on the CEO’s total compensation. In the year itself the compensation is 6.71% higher, the following year this effect is still 4.16%. Including diversification increases the effect even further. Diversifying acquisitions have a positive effect of 6.99 – 8.16% on the total compensation in comparison to focused acquisitions. This shows a compensation driven incentive for CEOs in pursuing acquisition activity. The effect of CEO seniority on acquisition’s announcement day returns is significantly positive with coefficients of .0308 and .0262. This implies that an increase in CEO age by 1% leads to a .03 percentage point higher return in a diversifying acquisition. When I divide age into three age groups my results show that younger CEOs outperform older CEOs in focused acquisitions by .57 percentage points. Middle aged CEOs do the best diversifying acquisitions. The two other age groups generate -.37 percentage points lower Cumulative Abnormal Returns in diversifying acquisitions.

This paper proceeds as follows. Section 2 covers the relevant literature and relates this research to it. Section 3 covers the hypotheses of this research. Section 4 explains the methodology and the variables used. Section 5 discusses the data and summary statistics. Section 6 covers the results of this paper. Section 7 covers robustness checks. In the last section a conclusion from the research will be completed.

2. Literature Review

This section covers the relevant literature. The section starts with showing why the CEO characteristic age is interesting as a topic of research. The literature shows that CEO decision making changes with age. After this, research regarding the effect of diversifying acquisitions on the

4_{See Custódio and Metzger (2013), Huang and Kisgen (2013), Jenter and Lewellen (2014), Malmendier and Tate}

(5)

5 acquisition returns is covered. Earlier literature shows that diversifying acquisitions have a negative acquisition return. The next part discusses two papers that focus on empire building as a corporate governance problem in specific. The two papers find contradicting results on empire building. The literature review concludes with several papers that focus on the relation between different CEO characteristics and acquisition returns. These papers show that CEO characteristics have a significant influence on the acquisition returns.

The effect of CEO age

McClelland et al (2012) research the relation between the CEO career horizon and future firm performance. The career horizon problem means that the horizon of the CEO is shorter than the horizon of shareholders. This can lead to older CEOs preferring short term risk averse strategies that personally benefit them but harm the shareholders. McClelland et al (2012) show that CEOs with a shorter career horizon perform worse than CEOs with a longer career horizon. However, this result depends on the level of CEO stock ownership. For lower levels of stock ownership there is no significant effect. For high levels of stock ownership the CEOs career horizon becomes a problem for shareholders. This is caused by higher levels of risk averseness. Serfling (2014) tests the relation between CEO age and risk taking behavior. His results show a negative relation between CEO age and stock return volatility. Serfling shows that older CEOs limit their risk by doing more diversified acquisitions and investing less in R&D. Serfling finds no support that his results are driven by the horizon problem. These papers (McClelland et al (2012) and Serfling (2014)) show that the decision making process by CEOs is affected by their age. The difference in decision making shows the relevance of testing CEO seniority in relation to acquisition returns.

Diversifying acquisitions

This research focuses on the difference between focused and diversifying acquisitions. There has been quite some research testing the effect of diversifying acquisitions on acquisition returns. Several papers (Berger and Ofek (1995), Lamont and Polk (2002), Morck et al (1990) and Hoechle et al (2012)) show that diversifying acquisitions lead to a loss of shareholder value. Morck et al (1990) show that there are 3 types of acquisitions that have systematically lower returns. These types are: diversifying acquisitions, acquisitions of high growth firms and acquisitions by poorly performing managers. They show that managers will overpay for targets if there are high private benefits. Morck et al (1990) see three different reasons why managers diversify their firms at the cost of shareholders. Firstly, managers can diversify their own risk this way. Secondly, managers enter new lines of business when shareholders prefer to shrink the firm. Thirdly, when the firm is performing poorly, managers enter industries they are good at. These papers relate to my research in their focus

(6)

6 on diversifying acquisitions. The papers show that diversifying acquisitions are generally value-destroying for shareholders while it may benefit executives. The differences in acquisition returns between focused and diversifying acquisitions create a relevant field of research. It can be interesting to see whether the relation between CEO seniority and acquisition returns differ between the types of acquisitions.

Empire building

Bertrand and Mullainathan (2003) focus on plant-level data instead of acquisitions. They test the relation between managerial preferences and corporate governance. Their results show that managers prefer the ‘quiet life’ instead of empire building. This result relates to this paper since this research focuses on empire building. Bertrand and Mullainathan find results contradicting empire building. It can be interesting to see if this research finds results for empire building or the so called ‘quiet life’.

Masulis et al (2007) test if corporate governance mechanisms affect the profitability of acquisitions. Using the GIM index they determine the level of Anti-Takeover Provisions. Based on the level of ATPs they determine the influence by the market for corporate control. Their results show that firms with more ATPs experience significantly lower acquisition returns. The result is economical significant, with an average loss of 56 million in shareholder value. Masulis et al (2007) argue that these managers are less affected by the market of corporate control, which leads to more value-destroying acquisitions. This research relates to this research in its focus on corporate governance issues in relation to acquisition returns. Masulis et al (2007) show that bad corporate governance has a negative effect on acquisition returns and leads to empire building. This research follows Masulis et al (2007) in taking value-destroying acquisitions as a proxy of empire building.

CEO characteristics and acquisition returns

Walters (2007) focuses on the relation between CEO tenure and acquisition returns. He tests this relation under two conditions: with and without a vigilant board. His results show a significant curvilinear relation between CEO tenure and acquisition returns. Acquisition performance increases from low to moderate levels of tenure. For longer tenure levels the relation with acquisition returns is negative. Walters finds an optimal level of tenure of 8.1 years. The negative relation of longer tenure levels with acquisition returns is insignificant in firms with a vigilant board. He argues that a new CEO is in a vulnerable position so will not pursue personal benefits. As the CEO gets more tenured he gets more entrenched and starts pursuing personal benefits. A vigilant board will prevent this from happening. The focus of the research by Walters (2007) is interesting for this paper. It can

(7)

7 act as comparison for the control variable tenure and shows that CEOs pursue personal benefits if possible.

Huang and Kisgen (2013) focus on another CEO characteristic. Their paper tests whether the executive’s gender causes a difference in financial decision making and acquisitions returns. The results show that male executives take more financial decisions such as acquisitions and issuing debt. Huang and Kisgen show that male executives have lower announcement returns than female executives. Female executives generate on average 2% higher acquisition announcement returns than male executives. Huang and Kisgen argue that this difference is caused due to overconfidence by male executives. They find that female executives place wider bounds on earnings estimates and exercise their options earlier. These two actions are seen as indicators of overconfidence. They show that male executives are more likely to be fired, which confirms that male overconfidence and not female risk averseness causes the difference in returns. Huang and Kisgen show the importance of CEO characteristics on acquisition returns. This research uses CEO gender as a control variable. It will be interesting to compare the outcome to this research.

Malmendier and Tate (2008) research the relation between CEO overconfidence and acquisition returns. They use two different proxies for overconfidence: overinvestment in the company and the CEO’s press portrayal. Their main result shows that overconfident CEOs overestimate the returns they can generate. Because of this overconfidence CEOs overpay and undertake value-destroying acquisitions. An overconfident CEO has a 65% higher chance of announcing an acquisition. The market reaction on these acquisitions is -90 basis points versus -12 basis points for non-overconfident CEOs. The effect is the strongest in diversifying acquisitions with access to internal financing. In focused deals the negative effect by overconfident CEOs is insignificant. According to Malmendier and Tate this shows that overconfident CEOs want to do extra deals even though these are bad quality deals. The focus on CEO characteristics is valuable for this research. Malmendier and Tate take diversification as a proxy for value-destruction at the cost of shareholders. This relates to the approach I take in this paper.

Custódio and Metzger (2013) test whether a CEO’s industry expertise affects the acquisition returns. Their results show a positive relation between industry expertise and acquisition returns. An acquiring CEO, with industry expertise in the target’s industry, has 1.2-2.0 percentage points higher abnormal returns than an inexperienced CEO. Custódio and Metzger make a distinction between value creation and value capturing as the two main channels for industry expertise. They show that their result is driven by value capturing instead of value creation. The experienced CEO has a better ability to capture more of the merger surplus. By being better bargainers these CEOs pay 7.5-9.75% lower premiums. This effect is mostly information-based. This means the positive effect is the largest when the information asymmetry is the highest. This research again shows the importance of the

(8)

8 CEO in acquisition returns. The research also shows the negative effect of diversifying acquisitions with a 2.2 percentage points lower CAR. This is relevant for the focus on diversifying acquisitions in this research. Furthermore, the method Custódio and Metzger use has similarities to the method this research uses. They construct similar regression models with interaction terms to test the effect of CEO experience.

Jenter and Lewellen (2014) test whether target CEO’s retirement preferences affect the pricing and propensity of acquisitions. They find a significant increase in the likelihood of a successful takeover bid when the CEO is close to retirement age. Just before the retirement age the likelihood of a successful bid is 4.4%. In the 64-66 age group this likelihood increases to 5.8%, which is a significant increase of 32%. This effect appears to be smaller in well governed firms. This may point towards agency problems as a channel for this effect. Their results also show that the acquisition premium and target announcement returns are insignificantly higher for CEOs at retirement age. This means that older CEOs are 32% more likely of selling their firm and do this without sacrificing acquisition premium. Jenter and Lewellen argue that agency problems cause the effect. However, they argue that younger CEOs might be reluctant to sell their firm even when it is value-increasing. They test this by showing that good governed firms see an increase in sales by young CEOs and just a slight increase by older CEOs. In good governed firms there are less agency problems and younger CEOs accept good bids as well. They show that the short term focus by CEOs near their retirement age is not always a disadvantage. This paper by Jenter and Lewellen shows the impact the CEO age can have on acquisition preferences. It relates to this research in the focus on the CEO age, although Jenter and Lewellen focus on target CEOs. An important aspect they show is the occurrence of managerial self-interest and how this changes among different CEO age groups.

The research closest to my research is by Yim (2013). Yim tests the relation between acquisition behavior and CEO age. The main result of her paper is a negative relation between CEO age and the acquisition propensity. She shows CEOs in the oldest age group (63-92) have a 28.5% lower probability of announcing an acquisition than a CEO in the lowest age group (27-48). She also looks at announcement day returns. She finds small insignificant results that younger CEOs have lower announcement day returns. The channel for her results is given by theory on compensation driven incentives. She shows that CEOs receive a significant increase in compensation following acquisitions. This positive effect is visible even three years later, accumulating to an increase in compensation of 23% over three years. The compensation incentive is strongest for younger CEOs since they have a longer career horizon left. The longer career horizon combined with the compensation incentive is the argumentation for her main result that younger CEOs are more likely to pursue acquisitions. The research by Yim relates to this research in a couple of ways. This paper also uses the theory on compensation driven incentives as argumentation for the results. This

(9)

9 research only focuses on the relation between age and acquisition returns, where Yim’s main focus is the propensity of doing acquisitions. A clear distinction between our papers is the separation of acquisition types that this research makes. I expect that this separation gives a better and significant result for the effect of CEO seniority on acquisition returns. This may add to the paper by Yim that younger CEOs not only increase their acquisition propensity but also do lower return acquisitions for the compensation incentive.

3. Hypotheses

This section discusses the hypotheses this research tests. The section begins with the hypotheses for the compensation incentive. The next part discusses the hypotheses for the effect of CEO seniority on acquisition returns.

This research starts by testing if there is an incentive for CEOs in pursuing acquisitions for their personal benefit. To provide this incentive the research looks at the effect of making an acquisition on the total compensation. Yim (2013) already shows a compensation incentive driven by acquisition activity. Following Yim (2013) and based on CEOs acting in their personal benefit the first hypothesis is as follows:

1. Acquisition activity has a significant positive effect on the CEOs total compensation.

This hypothesis can act as an incentive for CEOs to do acquisitions even when this is not in the shareholders’ best interest. This research focuses on the difference between diversifying and focused acquisitions. If these acquisitions generate different returns this should be visible in the compensation incentive as well. This leads to the next hypothesis:

2. Diversifying acquisitions have a significantly larger effect on the CEOs total compensation than focused acquisitions.

These two hypotheses focus on the channel for my further results. When a CEO has a compensation driven incentive to pursue acquisitions, he might act in his own benefit. I reason the compensation incentive is stronger for younger CEOs since they have a longer horizon to benefit.

This paper will continue by testing three hypotheses related to the effect of CEO seniority on (diversifying) acquisition returns. The research follows the approach by Masulis et al (2007) that value destroying acquisitions are a sign of empire building. Malmendier and Tate (2008) use a similar approach by taking diversifying acquisitions as a proxy for value-destruction at the shareholders’

(10)

10 costs. This research uses diversifying acquisitions as a proxy of empire building. Following Bertrand and Mullainathan (2003) I reason that CEOs nearing their retirement will start enjoying the ‘quiet life’. Younger CEOs might have a larger incentive to increase compensation due to their longer career horizon. They can benefit from the compensation longer because of the longer career horizon. This research argues that younger CEOs are driven by the compensation incentive and not by possible acquisition returns. This leads to the third hypothesis:

3. Acquisition’s announcement returns have a positive relation with age

This hypothesis corresponds to the research by Yim (2013). Yim finds an insignificant negative announcement return for younger CEOs. This hypothesis implies younger CEOs have a bigger tendency for empire building and the associated power, prestige and compensation. According to earlier research, diversification leads to a loss of shareholder value5. The empire building motives make the CEO indifferent to the lower shareholder value. Earlier hypotheses argue that diversifying acquisitions cause an even larger compensation incentive for CEOs. Due to the longer career horizon this might mean that younger CEOs are more incentivized to engage in value-destroying diversifying acquisitions than older CEOs. These assumptions lead to the fourth hypothesis:

4. Diversifying acquisition’s announcement returns have a positive relation with age.

If the compensation incentive is stronger for diversifying acquisitions the effect for focused acquisitions should be insignificant or significantly smaller. Opposing to the fourth hypothesis, the fifth hypothesis argues that focused acquisition announcement returns do not have a significant relation with CEO age:

5. Focused acquisition’s announcement returns are indifferent to CEO age

The results of the fifth hypothesis are compared to the earlier hypotheses. It will act as control group to the fourth hypothesis.

4. Methodology

This section discusses the method used to test the hypotheses. The section also explains the variables used in the regressions.

(11)

11

4.1 Research Method

The research begins with testing whether CEOs have a compensation incentive in pursuing acquisitions. The dataset to test the effect of acquisition activity on the CEO compensation differs from the dataset used for the effect of CEO seniority on acquisition returns. The dataset contains all CEOs of US listed firms in 1995-2014. The CEOs doing acquisitions will be subdivided into CEOs doing diversifying acquisitions and CEOs doing non-diversifying acquisitions. The following OLS regressions test the first two hypotheses:

1.

2.

To back up the reasoning that CEOs have a compensation incentive in pursuing diversifying acquisitions this research takes two steps. The research starts with testing whether CEOs that do acquisitions receive a larger compensation than CEOs that do not do acquisitions. The second regression tests if performing a diversifying acquisition leads to a larger compensation. This research uses event studies and OLS regressions to test the hypotheses on the relation between CEO seniority and acquisition returns. I follow earlier research on CEO characteristics and acquisition returns that use event studies and OLS6. For each acquisition a 41-day window will be used to derive the Cumulative Abnormal Returns. The main event time will be 1 day before and 1 day after the event. The event will be the announcement day of the acquisition. The following regression models look for short term return (market reaction):

1.

2.

The first regression functions as comparison with Yim (2013) and test the third hypothesis. The second regression is of importance in my research. The interaction term will test hypothesis four about the relation between CEO seniority and diversifying acquisition returns. This approach is similar to the approach followed by Custodio & Metzger (2013). The regression without the interaction term (dummy is zero) will test the effect of age on focused acquisitions. According to

6_{See Custódio and Metzger (2013), Huang and Kisgen (2013), Jenter and Lewellen (2014), Malmendier and Tate}

(12)

12 hypothesis four the interaction term is significantly positive in this regression. According to hypothesis five the variable Age in the second regression is insignificant.

4.2 Variables

The variables used in the regressions are similar for all the OLS models. The variable age is the logarithm of age. The logarithm is used to normalize the distribution since CEO age probably has a skewed distribution. The variable age will be a continuous variable. I will also use age groups in certain regressions. In this case terciles are constructed by age to see the effect of different age groups on the acquisition returns. The three age groups are dummies for young CEOs, which are CEOs in the age group 38-51; middle aged CEOs in the age group 52-58 and old CEOs in the age group 59-74. I expect younger CEOs to generate lower acquisition returns. Due to the compensation incentive in pursuing acquisitions they might care less about the acquisition returns. This leads to reason that a positive coefficient for the age variable is expected.

The variable Diversification is a dummy variable indicating a diversifying acquisition. The dummy is 1 for a diversifying acquisition and 0 for a non-diversifying acquisition. As indicator for diversifying acquisitions the research uses SIC codes provided by Compustat-CRSP, just as Serfling (2014). The SIC code is a 4-digit code indicating the industry of the firm. The constructed dummies are based on the first 3-digits and 2-digits of the SIC code. A difference in the first three or two digits indicates a diversifying acquisition. Both the 3-digits and 2-digits are used in regressions to be sure to cover the industry allocation properly. Using these two different indicators of diversifying acquisitions covers robustness issues this might bring. Earlier research shows a negative effect of diversifying acquisitions on the returns7. I expect a negative coefficient for the diversification dummies.

The interaction term Diversification*Log(Age) is constructed by multiplying the diversifying acquisition dummy with the age variable. This interaction term will show the combined effect of age and diversification on the Cumulative Abnormal Return. I expect that younger CEOs are more motivated by the compensation incentive to pursue acquisition activity. The compensation incentive may lead them to undertake value destroying diversifying acquisitions. Because of this a positive coefficient for the interaction term is expected.

To test the channel of the results two dummy variables are constructed: acquisition [t] and acquisition [t-1]. The dummy acquisition [t] is 1 if the CEO did an acquisition in that year. The acquisition [t-1] dummy is 1 if the CEO did an acquisition in the year before. Acquisition [t-1] is

(13)

13 included to show that the effect on compensation by acquisitions is not temporarily and still has influence a year later. Since I argue that there is a compensation incentive for CEOs in pursuing acquisitions I expect both coefficients to be positive.

As control variables I use several CEO characteristics as CEO tenure, gender and the amount of shares the CEO owns. Since this paper focuses on a specific CEO characteristic it is important to mitigate omitted variable bias. Other control variables are firm characteristics such as total assets, market-to-book ratio, leverage, EBITDA/Assets and Sales/Assets. Just as with the CEO characteristics, firm specifics can bias the results.

Looking at other event windows than the [-1,1] window covers robustness problems. Other windows will be [0,2], [-2,0] and [-2,2]. Inclusion of other event windows checks whether the selection of the event window influences the results. Besides applying all the control variables to the regressions industry and year fixed effects are added. Industry fixed effects are included since the firms in my sample range over many industries that have systematic differences in risk and performance. The fixed effects control for this. The same reasoning applies to the inclusion of year fixed effects. It controls for systematic differences in risk and performance over the multiple years in my sample. Following earlier research on the relation between CEO characteristics and acquisition returns I do not include firm fixed effects8. None of the earlier papers use firm fixed effects in the regressions testing the effect of CEO characteristics on acquisition returns. In the appendix one table including firm fixed effects is included to show the effects.

Following earlier research on the effect of CEO characteristics on acquisition returns a threshold of the acquisition size as robustness checks is used9. This threshold will be 5%, meaning that firms that are at least 5% of the market capitalization are used in this regression. This threshold is to test if there is a difference in the impact by small acquisitions and larger acquisitions.

5. Data & descriptive statistics

This section discusses the used databases and construction of the datasets. Furthermore, this section discusses the summary statistics of the datasets.

The data on acquisitions is from the Thomson One Database. I retrieve all acquisitions in the period 1995-2014 by a public acquirer. Execucomp provides data since 1995, which limits me from going back further. The end date of the research ensures a recent time period. Only firms from the United States acquiring another American firm are included. Taking only transactions that have a change of ownership gives 61,957 observations. After excluding transactions without a deal value

8_{See Custódio and Metzger (2013), Huang and Kisgen (2013) and Yim (2013)} 9_{See Malmendier and Tate (2008), Morck et al (1990) and Yim (2013)}

(14)

14 there are 33,062 observations. Into this dataset I merge basic quarterly firm fundamentals from CRSP/Compustat-Merged as EBITDA, leverage, size and market capitalization. This results in 20,514 unique observations. Execucomp provides data on CEO tenure, gender, age and stock ownership. I only keep executives with the title of CEO, which are 41,030 observations. Merging this with the dataset containing the acquisitions gives 9,221 observations matched on acquisition-CEO level. This dataset contains duplicates on CEO level and missing values on tenure and age. I complete this dataset by hand. I collect information about CEO age and tenure on Capital IQ and Bloomberg. These databases also provide me with information regarding the decision on removing duplicates. There are double CEO entries possible due to CEO turnover during the year of the acquisition. After finishing the dataset by hand there is a total amount of 8,310 observations on the CEO-acquisition level. This dataset is expanded to a 41-day window. To compute the Cumulative Abnormal Returns this dataset is completed with stock and market returns provided by CRSP. Observations that do not have a return on the announcement day of the acquisition are deleted. Eventually the dataset includes 7,967 unique observations on a CEO-year level.

I construct a different dataset to empirically test whether there is a compensation incentive for CEOs in pursuing acquisition activity. I begin with the dataset from Execucomp including all CEOs of US firms in the period 1995-2014. In this dataset the acquisition data from Thomson One is merged and all acquisitions that are not allocated to a CEO are dropped. This gives 43,941 observations. This dataset is completed by adding in annual fundamentals for all the firms from the CRSP/Compustat-Merged database. After that merge there are 42,845 observations left. After dropping duplicates there are 38,362 unique CEO-firm year level observations left. This dataset contains 6,517 acquisitions. This number is lower than in the other dataset due to the fact that in this dataset CEOs conducting multiple acquisitions in one year are represented only once. This is necessary to prevent double counting. In both datasets all variables are winsorized at 1% to control for outliers that might bias the results.

Table 1 provides the summary statistics of the dataset for the channel of the research. The summary statistics in table 1 shows differences between acquiring firms, non-acquiring firms and their CEOs. The table shows significant differences in the means of the samples. Panel B shows firm and CEO characteristics of the full sample and separately for acquiring firms. The means and medians of the variables of acquiring firms are t-tested against the full sample means and medians. Panel C shows the firm and CEO characteristics of firms that undertake diversifying acquisitions. The means and medians of the variables in panel C are t-tested against the means and medians of acquiring firms in panel B.

Acquisition values range from $7,000 to $130,298 billion with a mean of $677 million. Market capitalization has an average of $6,297 million in the full sample. For acquiring firms this average is

(15)

15 $8,655 million. This is significantly different at 1% from the full sample average market capitalization. This shows that firms doing acquisitions are in general larger than non-acquiring firms. The average market capitalization is higher for firms doing diversifying acquisitions. An average of $9,605 million for firms doing acquisitions with a 3-digit different SIC code and $10,225 million for acquisitions with a 2-digit different SIC code. Both means are significantly different at 1% from the full sample mean of acquiring firms. This shows that firms doing diversifying acquisitions are on average larger than the average firm in the full sample of acquiring firms. The Tobin’s Q is 1.32 in the full sample and 1.48 for the sample of acquiring firms; this is significantly different at 1%. For firms doing diversifying acquisition the average Tobin’s Q is 1.32 and 1.35. These are significantly different at 1% and 5% from the total sample of acquiring firms. Total assets of the full sample are $10,471 million. For acquiring firms this average is $12,898. For diversifying firms these averages are $14,728 and $15,115 million. The total assets show the same pattern as the market capitalization: assets are larger for firms doing acquisitions. The firms in the full sample have average sales of $4,918 million. Acquiring firms in this sample have average sales of $5,973 million, which is significantly higher at 1% than the full sample average. Diversifying firms have sales averages of $7,163 and $7,773 million. Both are significantly different from the sample of acquiring firms at 1%. Table 1 shows a clear pattern for all the size proxies (market capitalization, assets and sales). Acquiring firms tend to be larger than non-acquiring firms. According to the used size proxies diversifying firms tend to be larger than focused firms. For EBITDA the full sample has an average of $754 million. For acquiring firms this average is $984 million which is significantly higher at 1%. For firms doing diversifying acquisitions the average is $1,104 million, which is significantly different at 1%. 2-Digit diversifying firms have an average EBITDA of $1,218 million, which is significantly different at 1%. The mean EBITDA shows that acquiring firms have a higher average EBITDA than non-acquiring firms. This makes sense since firms need free cash flows to finance acquisitions. Leverage is 0.70 for firms in the full sample. Acquiring firms have an average leverage of 0.86, which is not significantly different from firms in the full sample. Firms doing 2-digit diversifying acquisitions have a leverage of 0.53. The full sample has an average CEO age of 56.09 years. This is 55.52 for the acquiring firms, which is significantly lower at 1%. CEOs doing 3-digit diversifying acquisitions have an average age of 55.97, which is significantly higher than the average age in the sample of acquiring firms. CEOs that do 2-digit diversifying acquisitions have a mean of 54.77, which is not significantly different from the full sample of acquiring firms. The CEO tenure in the full sample averages 6.78 years. For acquiring firms this is 6.85 years. CEOs that do diversifying acquisitions have an average tenure of 6.68 and 6.66. In all the cases the CEO tenure is not significantly different from the full sample. In the full sample an average of 2.5% of the CEOs is female. For the sample of acquiring firms there are on average 1.9% female CEOs, which is significantly lower than the full sample average at 1%. Stock ownership is 3.00% on average

(16)

16 in the full sample, which is slightly higher compared to the 2.8% found by Kim and Lu (2011). For acquiring CEOs the stock ownership is on average 2.59%, which is significantly lower than in the full sample. For CEOs doing diversifying acquisitions the stock ownership by CEOs is 2.33% and 2.36% respectively. On average CEOs that do acquisitions own less of the firm. The average total compensation received by the CEOs in the full sample is $4.6 million. For acquiring CEOs this is an average of approximately $5.7 million. This is significantly higher at 1% than the full sample CEO average compensation. For CEOs doing acquisitions with a 3-digit different SIC code the average compensation is $6.0 million. CEOs doing acquisitions with a 2-digit different SIC code earn an average compensation of $6.2 million. This is higher than the acquiring firm sample with 1% significance.

Table 1

Summary statistics of the dataset used to test the compensation incentive. Observations are at the CEO-firm level in the period 1995-2014. Panel A shows the value of the included transactions as announced in Thomson One. Panel B shows specific CEO and firm characteristics for the full sample and for acquiring firms in the sample. Market capitalization is the value of outstanding shares multiplied by the closing price. Tobin’s Q is the market capitalization divided by total assets. Total Assets is the book value of assets. Sales is the total value of sales for that year. EBITDA is included as proxy for free cash flows. Leverage is total long term debt divided by common equity. CEO age is the age in that year. CEO tenure is the time the executive is in his position at the current firm. Female is a dummy variable that is 1 in case of a female CEO. Stock ownership shows the amount of stocks the CEO owns in the firm. Total compensation is the compensation including bonuses the CEO received. Panel C shows the same information as panel B. However, panel C includes only acquiring firms and separates them on the kind of acquisition. 3-Digit SIC means the merging firms are different in the first three digits of the SIC code. 2-Digit SIC means a difference in the first two digits of the SIC codes. The means and medians of acquiring firms are compared to the means and medians of the full sample. The means and medians in panel C are compared to the means and medians of acquiring firms in panel B. Significance levels as follows: * at 10%, ** at 5% and *** at 1%.

Panel A: Deal Specifics

N Mean Median Std. Dev. Min Max Value of Transaction 6,519 677.48 88.75 3,877 0.007 130,298

Panel B: Firm and CEO characteristics

Full Sample Acquiring firms

N Mean Median Std. Dev. Min Max N Mean Median Std. Dev. Min Max Market Capitalization 38,345 6,297 1,661 12,250 2.78 61,068 6,519 8,655*** 2,276*** 15,196 3.22 61,068 Tobin's Q 38,345 1.32 0.86 2.00 0.002 105 6,519 1.48*** 0.95*** 2.51 0.007 82.17 Total Assets 38,363 10,471 1,999 24,142 3.79 138,898 6,519 12,898*** 2,516 27,933 9.28 138,898 Sales 38,359 4,918 1,390 9,307 0.00 47,979 6,519 5,973*** 1,392 11,123 0.00 47,979 EBITDA 30,991 753.61 178.19 1,566 -78.50 8,134 5,432 984.08*** 217.60 1,874 -78.50 8,134 Leverage 38,231 0.70 0.41 31.90 -1,905 3,527 6,501 0.86 0.43 16.92 -368.41 845.02 CEO Age 37,570 56.09 56 7.28 39 76 6,377 55.52*** 55*** 7.22 39 76 CEO Tenure 33,470 6.78 5 6.88 0 33 5,718 6.85 5 6.55 0 33 Gender 38,364 0.025 0 0.16 0 1 6,519 0.019*** 0 0.14 0 1 Stock Ownership 22,744 3.00 0.63 6.16 0 35.2 3,766 2.59*** 0.65 5.40 0 35.2 Total Compensation 38,141 4,627 2,625 5,754 151.67 36,649 6,488 5,739*** 3,203*** 7,063 151.67 36,649

Panel C: Firm and CEO characteristics of firms doing diversifying acquisitions

Diversifying Acquisitions (3-digit SIC) Diversifying Acquisitions (2-digit SIC)

N Mean Median Std. Dev. Min Max N Mean Median Std. Dev. Min Max Market Capitalization 3,523 9,605*** 2,640 16,030 3.22 61,068 2,717 10,225*** 2,741 16,653 3.22 61,068 Tobin's Q 3,523 1.32*** 0.91 2.05 0.007 48.69 2,717 1.35** 0.93 2.22 0.007 48.69 Total Assets 3,523 14,728*** 2,965 30,274 10.84 138,898 2,717 15,115*** 2,965 30,923 10.84 138,898 Sales 3,523 7,163*** 1,766 12,228 0.00 47,979 2,717 7,773*** 1,764 13,024 0 47,979 EBITDA 2,801 1,104*** 260.7 1,959 -78.50 8,134 2,163 1,218*** 274.26 2,074 -78.50 8,134 Leverage 3,520 0.64 0.47 6.15 -116.68 155.60 2,715 0.53 0.46 4.03 -116.68 64.78 CEO Age 3,428 55.97*** 56*** 7.17 39 76 2,642 55.77 56*** 7.24 39 76 CEO Tenure 3,054 6.68 5 6.45 0 33 2,358 6.66 5 6.49 0 33 Gender 3,523 0.023 0 0.15 0 1 2,717 0.025* 0 0.16 0 1 Stock Ownership 1,979 2.33* 0.60 5.16 0 35.2 1,540 2.36 0.60 5.18 0 35.2 Total Compensation 3,506 6,029* 3,420 7,228 151.67 36,649 2,704 6,232*** 3,505* 7,433 151.67 36,649

(17)

17 Table 2 shows the summary statistics of the dataset for the main research. This dataset is constructed with the use of quarterly fundamentals provided by CRSP/Compustat. The use of quarterly data instead of yearly data might lead to different numbers compared to the summary statistics in table 1. The summary statistics in table 2 also excludes the transactions where the stock returns on the announcement day are missing. Table 2 provides comparison between the full sample of acquiring firms and firms doing 2-digit SIC code diversifying acquisitions. In panel A the deal specifics show a total of 7,967 acquisitions with an average transaction value of $509 million. Of the 7,967 acquisitions there are 3,414 diversifying acquisitions. The average transaction value for diversifying acquisitions is $387 million, which is significantly lower at 1% than the full sample mean. The full sample of firms has an average market capitalization of $9,303 million, whereas the diversifying acquisitions have a market capitalization of $11,337 million. This is significantly larger at 1% than the average of the full sample. In the full sample of acquiring firms the average total assets is $11,320 million. Firms doing diversifying acquisitions have an average of $13,493 million in total assets. This average is different from the full sample average at 1% significance level. The averages of market capitalization and total assets show that diversifying firms tend to be larger than focused firms. Tobin’s Q for the full sample of acquiring firms is 1.77 on average. For diversifying acquisitions this average is 2.38. Leverage of acquiring firms is 0.64 in my sample. The average of diversifying firms is 0.52; this is not significantly different from the full sample. The EBITDA of acquiring firms is $268 million, compared to $346 million for diversified acquiring firms. This average of diversified acquiring firms is significantly higher at 1% than the full sample average. The CEO age in the full sample has an average of 54.41 years, compared to 54.75 years for CEOs of firms doing diversifying acquisitions. The average CEO age of firms doing diversifying acquisitions is significantly higher at 5%. This shows that CEOs of diversifying firms are on average older than CEOs of focused firms. CEO tenure is on average 7.10 years for the full sample and 6.85 years for the subsection of diversifying firms. In the full sample 1.6% of the CEOs are female against 1.9% female CEOs for diversifying firms. Stock ownership is on average 2.72% in the full sample and 2.47% for the CEOs of firms doing diversifying acquisitions. Both averages are slightly below the amount of stockownership that Kim and Lu (2011) show. The average total compensation for CEOs of acquiring firms is $6.9 million, compared to an average of $7.7 million for CEOs doing diversifying acquisitions. This difference is significantly larger at a 1% significance level, which shows that CEOs in diversified firms on average earn a higher compensation. The average Cumulative Abnormal Return in the [-1,1] window is 0.4%, compared to an average of 0.35% for firms doing diversifying acquisitions. This average is not significantly different than the full sample average. Table 2 shows that there are significant differences between the full sample and the sample containing firms that do diversifying acquisitions. Most of these differences are in firm characteristics, as is already shown in table 1.

(18)

18 Table 2

Summary statistics of the dataset to test the effect of CEO seniority on announcement day returns. Observations are at acquisition-CEO-year level in the period 1995-2014. Panel A shows the value of the included transactions as announced in Thomson One. Panel B shows specific CEO and firm characteristics for the full sample and separate for one of the dummies indicating a diversifying acquisition. The notion 2-digit SIC means the merging firms are different in the first two SIC digits. Market capitalization is the value of outstanding shares multiplied by the closing price. Total Assets is the book value of assets. Tobin’s Q is the market capitalization divided by total assets. Leverage is total long term debt divided by common equity. EBITDA is included as proxy for free cash flows. CEO age shows the age at the time of the acquisition. CEO tenure is the time the executive is in his position at the current firm. Female is a dummy variable that is 1 in case of a female CEO. Stock ownership shows the amount of stocks the CEO owns in the firm. Total compensation is the compensation including bonuses the CEO received. CAR is the Cumulative Abnormal Return in the [-1,1] window. The diversifying acquisitions included in the summary statistics are compared to the full sample in order to see if they are significantly different. Significance levels as follows: * at 10%, ** at 5% and *** at 1%.

6. Results

This section presents the regression results to test the hypotheses. The section starts with testing whether CEOs have compensation incentives in pursuing acquisitions. Section 6.2 tests the effect of CEO seniority on the (diversifying) acquisition returns.

6.1 Compensation incentive for acquisition activity

As discussed in the methodology section this procedure consists of two steps. The first step is testing if acquisition activity has a significant effect on the total CEO compensation. The second step tests if there is a difference in compensation for CEOs pursuing diversifying or focused acquisitions.

Table 3 shows the Ordinary Least Squares regression results for the effect of acquisition activity on CEO compensation. Columns 1-6 show the regressions on the full sample of CEOs in US listed firms in the period 1995-2014. The regressions in columns 1-3 are without control variables and fixed effects. The acquisition [t] dummy indicates acquisition activity in the current year. The acquisition [t-1] dummy indicates acquisition activity in the year before. Column 1 shows the uncontrolled effect of acquisition activity on the total compensation. The coefficient for the

Panel A: Deal Specifics

Full Sample Diversifying Acquisitions (2-digit SIC)

N Mean Median Std. Dev. Min Max N Mean Median Std. Dev. Min Max Value of Transaction 7,967 509.04 78.5 2,502 0.01 89,168 3,414 387.02*** 74 1,584 0.01 40,298

Panel B: Firm and CEO characteristics

Full Sample Diversifying Acquisitions (2-digit SIC)

N Mean Median Std. Dev. Min Max N Mean Median Std. Dev. Min Max Market Capitalization 7,967 9,303 2,306 15,862 10.76 57,635 3,414 11,337*** 2,715 17,913 12.55 57,635 Assets 7,967 11,320 2,342 23,985 7.60 124,291 3,414 13,493*** 2,802 26,755 12.09 124,291 Tobin's Q 7,967 1.77 1.08 3.55 0.02 89.34 3,414 1.51*** 1.01 2.38 0.02 69.77 Leverage 7,888 0.64 0.41 15.21 -357.71 845.02 3,384 0.52 0.45 6.926 -357.71 80.57 EBITDA 6,529 268.31 58.67 497 -37.5 1,968 2,592 345.98*** 72.75 564.74 -37.5 1,968 CEO Age 7,967 54.41 54 7.36 38 74 3,414 54.75** 55*** 7.42 38 74 CEO Tenure 7,967 7.10 5 6.66 0 33 3,414 6.85* 5 6.60 0 33 Female 7,967 0.016 0 0.12 0 1 3,414 0.019 0 0.14 0 1 Stock Ownership 4,784 2.72 0.77 5.26 0.003 30.5 1,985 2.47* 0.69 4.98 0.003 30.5 Total Compensation 7,930 6,955 3,468 10,288 190.4 65,589 3,396 7,701*** 3,764 11,191 190.42 65,589 CAR 7,967 0.004 0.002 0.054 -0.170 0.184 3,414 0.0035 .002 0.049 -0.170 0.184

(19)

19 acquisition [t] dummy is .1772 and is significant at 1%. The coefficient for the acquisition [t-1] dummy is .2507 and significant at 1% as well. Economically this shows that acquisition activity leads to a 17.72% higher total compensation in the same year. The next year this effect is even higher with 25.07%. The regressions in columns 2-3 include the effect of diversification on the total compensation. The acquisition [t] dummy in this case shows the effect of a focused acquisition (if the diversification dummy is zero). Including diversification lowers the coefficient of acquisition [t] to .0982 in column 2 and .1152 in column 3. However, it remains significant at 1%. A year later the effect of acquisition activity on the total compensation remains high. The coefficient of acquisition [t-1] stays approximately .25 and significant at 1%. The effect of diversification is significant at 1% as well. Column 2 looks at 3-digit SIC diversifying acquisitions and shows a coefficient of .1462. This means a 14.62% higher total compensation for CEOs doing diversifying acquisitions than CEOs doing focused acquisitions. Column 3 shows the effect of 2-digit diversifying acquisitions on the total compensation. The coefficient of the diversification dummy is .1492 in this regression. This means a higher total compensation of 14.92% in comparison to a focused acquisition. The uncontrolled regressions show that both focused and diversifying acquisitions have a significant effect on the CEO’s total compensation. The effect of a diversifying acquisition is economically larger than that of a focused acquisition.

Columns 4-6 include several control variables that could affect the CEOs total compensation. These control variables are included to prevent omitted variable bias. The regressions in columns 4-6 also include industry and year fixed effects. In column 4 the coefficient for the acquisition [t] dummy of .0671 is significant at 1%. This shows that after including control variables acquisition activity still has a significantly large effect on the total compensation. From an economical perspective the coefficient implies a 6.71% higher compensation than the compensation of a non-acquiring CEO. The coefficient for the acquisition [t-1] dummy is .0416. Statistically this has a 5% significant effect on the CEOs total compensation. Looking at the economic perspective this means the CEO gets a 4.16% higher total compensation the year after an acquisition. The regression results from column 4 in table 3 show that there might be a compensation driven incentive for CEOs to pursue acquisitions. If this is the case it could mean that CEOs might pursue acquisitions even if this is not in the best interest of the firm’s shareholders. The Market-to-Book ratio has a significant large positive effect on the total compensation at a 1% significance level. Assets, Sales/Assets and the EBITDA/Assets ratio have a large positive effect as well. All three are significant at 1%. The leverage ratio has a slightly significant negative effect on the compensation level. CEO tenure has a positive effect on the compensation, which is significant at 1%. The level of CEO stock ownership has a negative effect on the total compensation, which is significant at 1%. Higher stock ownership and lower total compensation fits in the policy of tying the CEOs incentives to that of the firm’s shareholders.

(20)

20 Unexpected is that age and CEO gender do not have a significant effect on the CEO compensation. In general these variables have an influence on the total compensation.

The regression in column 5 of table 3 shows the effect of diversifying and focused acquisitions on the total CEO compensation. The coefficient of the 3-digit diversification dummy has a value of .0816, which is significant at 1%. This means that an acquisition that differs on the first three SIC code digits results in an 8.16% higher total compensation for the CEO. This is a statistical and economical significantly large increase. The inclusion of the diversification dummy leads to an insignificant coefficient for the acquisition [t] dummy. This means focused acquisitions do not significantly affect the CEO’s total compensation in this regression and might imply that the result from column 4 is driven by the effect of diversifying acquisitions. The dummy acquisition [t-1] is still significant at 5% with a coefficient of .0417. This shows that the year after the acquisition the CEO receives a 4.17% higher compensation than his non-acquiring counterpart. The control variables show a similar pattern as in column 4.

The regression in column 6 of table 3 is almost identical to the regression in column 5. The difference is the dummy variable indicating a diversifying acquisition. In this regression a diversifying acquisition is marked as such when the first two SIC code digits differ. The coefficient of the 2-digit diversification dummy is .0699, which is significant at 5%. From an economical perspective this coefficient means a 6.99% higher total compensation if a diversifying acquisition completed. The acquisition [t] dummy is significant at 5%. The coefficient of .0410 implies a 4.10% higher compensation in the year of the focused acquisition. The dummy acquisition [t-1] is significant at 5% with a coefficient of .0415. This means a 4.15% higher total compensation the year after the acquisition activity. The control variables in column 6 also show the same pattern as the control variables in columns 4 and 5.

The regressions in table 3 show that there are personal incentives for CEOs in pursuing acquisitions. Column 4 shows that CEOs pursuing acquisitions receive a 6.71% higher compensation that year. The year after the acquisition this effect is still 4.16%. These economically large increases might lead to actions by CEOs that are not beneficial for shareholders. The regressions in columns 5-6 show that the effect is possibly driven by diversifying acquisitions. The dummy acquisition [t] gets lower and less significant. The diversification dummies are larger and significant. The compensation for a CEO that completes diversifying acquisitions is 6.99% to 8.16% higher than that of a CEO doing focused acquisitions. This might mean that CEOs that act in their own benefit might prefer a diversifying acquisition over focused acquisitions. This kind of acquisition increases their personal gains substantially. The results from table 3 show support for the first two hypotheses in section 3. CEOs increase their compensation by completing acquisitions, which supports hypothesis one. Looking into the type of acquisition the results show that diversifying acquisitions have a significantly

(21)

21 large effect over focused acquisitions. This supports hypothesis two that the effect of diversifying acquisitions on the CEO’s total compensation is even larger.

Table 3

This table shows the effect of acquisition activity on the total CEO compensation. The dataset contains all CEOs from US firms in the period 1995-2014. In all the regressions the logarithm of the total compensation is the dependent variable. Columns 1-3 show the effect of acquisition activity on the total compensation of the CEO without control variables and fixed effects. The acquisition variables are dummies indicating an acquisition in that year or in the previous year (Acquisition [t-1]). The diversification variables are dummies indicating a diversifying acquisition. The notion of 3-digit and 2-digit refers to the method of allocating diversification with the SIC codes. Columns 3-6 include fixed effects and the control variables age, Market-to-Book Ratio, total assets, leverage, EBITDA/Assets, Sales/Assets, tenure, stock ownership and a female dummy. Total Assets is the book value of assets. The Market-to-Book-Ratio is the market capitalization divided by total assets. Leverage is total long term debt divided by common equity. EBITDA is included as a ratio over total assets. Sales is included as a ratio over total assets. CEO tenure is the time the executive is in his position at the current firm. Stock ownership shows the amount of stocks the CEO owns in the firm. Female is 1 if the CEO is a female, 0 if the CEO is a male. If required the control variables are converted to logarithms. All regressions are robust to control for heteroskedasticity. Significance levels as follows: * at 10%, ** at 5% and *** at 1%.

Log(TotalComp) Log(TotalComp) Log(TotalComp) Log(TotalComp) Log(TotalComp) Log(TotalComp)

(1) (2) (3) (4) (5) (6) Acquisition [t] .1772*** .0982*** .1152*** .0671*** .0267 .0410** (.0154) (.0218) (.0192) (.0165) (.0237) (.0209) Acquisition [t-1] .2507*** .2503*** .2496*** .0416** .0417** .0415** (.0167) (.0167) (.0167) (.0177) (.0177) (.0177) Diversification (3-digit) .1462*** .0816*** (.0278) (.0307) Diversification (2-digit) .1492*** .0699** (.0281) (.0311) Log(Age) -.0736 -.0762 -.0744 (.0609) (.0609) (.0609) Market-to-Book Ratio .0307*** .0308*** .0308*** (.0090) (.0090) (.0090) Log(Assets) .4373*** .4372*** .4371*** (.0062) (.0062) (.0062) Leverage -.0002* -.0002* -.0002* (.0001) (.0001) (.0001) EBITDA/Assets .7401*** .7398*** .7401*** (.0861) (.0861) (.0861) Sales/Assets .0397*** .0396*** .0395*** (.0139) (.0139) (.0139) Tenure .0032*** .0032*** .0032*** (.0011) (.0011) (.0011) Stock Ownership -.0279*** -.0278*** -.0278*** (.0016) (.0016) (.0016) Female .0024 .0005 .0010 (.0370) (.0369) (.0370) Constant 7.80*** 7.80*** 7.80*** 4.13*** 4.14*** 4.13*** (.0063) (.0063) (.0063) (.3307) (.3306) (.3306)

Year Fixed Effects No No No Yes Yes Yes

Industry Fixed Effects No No No Yes Yes Yes

N 38,141 38,141 38,141 17,320 17,320 17,320

(22)

22

6.2 CEO seniority and acquisition returns

Following the results from table 3 the next section will test the main hypotheses. Table 4 shows the results from the Ordinary Least Squares regressions with the Cumulative Abnormal Return as dependent variable. The data in table 4 only contains firms that did an acquisition. This table uses the variable age as a continuous variable to see the effect on the Cumulative Abnormal Return. Columns 1-3 show uncontrolled regressions with the age variable, diversification dummies and the dependent variable CAR. In column 1 the coefficient of the variable age has a negative coefficient of -.0096, which is significant at 5%. This means that an increase of 1% in CEO age leads to a decrease of -.0096 percentage points in the Cumulative Abnormal Return. Columns 2-3 include the effect of diversification and the interaction term between age and diversification. Age has a significant negative effect on the CAR. The coefficient increased with the inclusion of the diversification dummies to -.0146 and -.0133. As shown by earlier research, diversification has a negative effect on the acquisition return10. The coefficients have a negative effect of -.0365 up to -.0381. However, these are insignificant. The interaction term in columns 2-3 has positive coefficients of .0095 and .0089. Both are insignificant.

Columns 4-6 show the same regressions with control variables and fixed effects. The variable age is negative but not significant anymore in these regressions. In column 4 this means that age does not have a significant effect on the acquisition return. In columns 5-6 this means that the CEO age does not have a significant effect on focused acquisition returns. The interaction term is significantly positive at 5% for 3-digit diversifying acquisitions. The coefficient of .0308 shows that a 1% increase in age leads to a .031 percentage point increase in CAR. The interaction term including 2-digit diversifying acquisitions is significant at 10% with a coefficient of .0262. This coefficient shows that a 1% increase in age leads to a .026 percentage point higher diversifying acquisition return. The coefficients of the interaction term show that older CEOs do significantly better diversifying acquisitions. The regressions in columns 5-6 also show a significantly negative relation between the diversification dummies and announcement day returns. The negative effect ranges from -.1069 until -.1264 percentage points on the CAR. This shows an economically large negative effect of diversifying acquisitions on the CAR of -10 to -12 percentage points. The Market-to-Book ratio, total assets and female dummy show a significant relation with the Cumulative Abnormal Return. The Market-to-Book ratio shows that firms with a high market valuation compared to their book value generate higher returns around the announcement day of an acquisition. Higher total assets lead to a significantly lower acquisition return. Furthermore, just as Huang and Kisgen (2013) my research

(23)

23 shows higher returns for female CEOs. Since the other control variables are insignificant I cannot discuss their implications.

Columns 7-9 only include acquisitions where the transaction value is at least 5% of the acquirer’s market capitalization. Earlier research argues that transactions below this 5% threshold might be too small to have an effect on the firm11. The smaller acquisitions could diminish the significance of the variables of interest. Columns 7-9 show that in my regressions the results are not particularly driven by larger or smaller acquisitions. The coefficients of the interaction term and diversification dummy are almost similar to columns 4-6. However, the coefficients are not significant anymore due to the increased standard errors. This may be caused by the lower number of observations due to the size constraint on the transaction value. The control variables show a similar pattern as in columns 4-6, except for the female dummy which is not significant in these regressions. Included in the appendix is table 8. This table shows the same regressions as columns 4-9 of table 4 with firm fixed effects. As discussed in the methodology this research follows earlier research in not including firm fixed effects in my main regressions12. As table 8 in the appendix shows, including firm fixed effects leads to insignificant variables except for the Market-to-Book ratio. However, the regressions do have a very high R-squared. The interaction term and diversification dummy do have a similar sign as the regression of table 4. The variable age is positive in some regressions and negative in others.

The results from table 4 show support for the hypotheses. The third hypothesis states:

Acquisition’s announcement returns have a positive relation with age. The regression in column 1

finds results contradicting this hypothesis, with a negative coefficient for the logarithm of age. However, after including control variables and fixed effects the effect was not significant anymore. The coefficient did remain negative. The results show an insignificant negative relation between acquisition returns and age. Table 4 does not show support for the first hypothesis. The fourth hypothesis states: Diversifying acquisition’s announcement returns have a positive relation with age. The interaction term between the age variable and diversification dummy shows result supporting this hypothesis. The regressions with control variables show significant results to support a positive relation between age and diversifying acquisition returns. Hypothesis five states: Focused

acquisition’s announcement returns are indifferent to CEO age. I interpret the results as if the

diversification dummies are zero. The age variable is the only variable of interest in this case. In the uncontrolled regressions the age variable shows a significant negative coefficient. In columns 4-9 the coefficient is not significant anymore when control variables and the 5% threshold are included. This

11_{See Malmendier and Tate (2008), Morck et al (1990) and Yim (2013)} 12_{See Custódio and Metzger (2013), Huang and Kisgen (2013) and Yim (2013)}

(24)

24 supports the fifth hypothesis of this research that there is no significant relation between focused acquisition returns and CEO age.

Table 4

This table shows the effect of CEO seniority on the Cumulative Abnormal Returns around the acquisition’s announcement day. The dataset contains all acquisitions by US listed firms in the period 1995-2014. In all the regressions the CAR, or Cumulative Abnormal Return, is the dependent variable. The CAR is constructed in an event window of [-1,1]. The event is the day of the acquisition’s announcement. Columns 1-3 show the effect of CEO age on the CAR without control variables and fixed effects. The age variable is the logarithm of CEO age and is included as a continuous variable. The dummy variables diversification (3-digit) and diversification (2-digit) show the effect of a diversifying acquisition on the CAR. The notion of 3-digit and 2-digit refers to the method of allocating diversification with SIC codes. The interaction terms of the variable Age and the Diversification dummy show the combined effect on the CAR. Columns 3-9 include fixed effects and the control variables Market-to-Book ratio, total assets, leverage, EBITDA/Assets, tenure, stock ownership and a female dummy. Total Assets is the book value of assets. The Market-to-Book-Ratio is the market capitalization divided by total assets. Leverage is total long term debt divided by common equity. EBITDA is included as a ratio over total assets. CEO tenure is the time the executive is in his position at the current firm. Stock ownership shows the amount of stocks the CEO owns in the firm. Female is 1 if the CEO is a female, 0 if the CEO is a male. If required the control variables are converted to logarithms. The regressions in column 7-9 include only acquisitions larger than 5% of the acquirer’s market capitalization. All regressions are robust to control for heteroskedasticity. Significance levels as follows: * at 10%, ** at 5% and *** at 1%.

CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1] CAR [-1,1]

(1) (2) (3) (4) (5) (6) (7) (8) (9) Diversification*Log(Age) (3-digit) .0095 .0308** .0304 (.0097) (.0151) (.0282) Diversification*Log(Age) (2-digit) .0089 .0262* .0260 (.0094) (.0155) (.0297) Diversification (3-digit) -.0381 -.1264** -.1276 (.0387) (.0605) (.1136) Diversification (2-digit) -.0365 -.1069* -.1090 (.0378) (.0623) (.1197) Log(Age) -.0096** -.0146* -.0133** -.0005 -.0146 -.0097 .0008 -.0129 -.0078 (.0048) (.0075) (.0067) (.0090) (.0114) (.0106) (.0170) (.0204) (.0192) Market-to-Book Ratio .0011*** .0011*** .0011*** .0016*** .0015*** .0016*** (.0004) (.0004) (.0004) (.0006) (.0006) (.0006) Log(Assets) -.0016** -.0015* -.0016* -.0042** -.0042** -.0042** (.0008) (.0008) (.0008) (.0018) (.0018) (.0018) Leverage .00003 .00002 .00002 .00003 .00002 .00002 (.00002) (.00002) (.00002) (.00004) (.00004) (.00004) EBITDA/Assets .0771 .0802 .0794 .0712 .0701 .0697 (.0548) (.0549) (.0549) (.1052) (.1064) (.1062) Tenure .0000 .00002 .00001 .00006 .00007 .00006 (.0002) (.0002) (.0002) (.0003) (.0003) (.0003) Stock Ownership .0002 .0002 .0002 .00009 .0001 .0001 (.0002) (.0002) (.0002) (.0005) (.0005) (.0005) Female .0162* .0163* .0163* .0234 .0242* .0231 (.0097) (.0097) (.0097) (.0145) (.0143) (.0144) Constant .0423** .0626** .0575** .0108 .0690 .0487 .0288 .0885 .0673 (.0191) (.0301) (.0268) (.0370) (.0461) (.0429) (.0710) (.0833) (.0787)

Year Fixed Effects No No No Yes Yes Yes Yes Yes Yes

Industry Fixed Effects No No No Yes Yes Yes Yes Yes Yes

Only acquisitions > 5% No No No No No No Yes Yes Yes

N 7,967 7,967 7,967 3,844 3,844 3,844 1,693 1,693 1,693

The effect of CEO seniority on diversifying acquisition returns

MSc Business Economics, Finance track

Master Thesis

The effect of CEO seniority on diversifying acquisition returns

Ward van den Boorn 6062768

July 2015

Table of Contents

1.

1.

2.