The impact of closely held shares on IPO firm’s value:
evidence from the Netherlands, United Kingdom and Belgium
(1999-2015)
Amsterdam Business School
Name Chantal Raap
Student number 10542876
Program Economics & Business
Specialization Finance & Organization
Number of ECTS 12
Supervisor Ilko Naaborg
Completion 11-02-2016
Abstract
The central question of this study involves the relationship between anti-takeover provisions and IPO firm value. This research is about the effect of closely held shares on the firm value at the initial public offering date in the Netherlands, United Kingdom and Belgium over the years 1999-2015. Closely held shares are shares held by shareholders directly affiliated with management. In the literature it is described that takeover mechanisms are frequently used while taking the firm to public. Companies that have a larger amount of closely held shares are more resistant to hostile takeovers. The research hypothesizes that closely held shares have a contrary effect on the firm’s value at the initial public offering date. This will be tested using multiple OLS regression analyses. Consistent with this hypothesis, this research reports that the percentage of closely held shares has a negative sign on IPO firm’s value. Although not significant for the combination of countries, there is a negative significant effect on IPO firm´s value in the United Kingdom. Moreover, a robustness check using different time periods provides evidence for a negative impact of closely held shares on IPO firm value in times without economic depression. Concluding, this study finds that managers entrench themselves at the cost of owners. Company owners should have a more tightened corporate governance structure to mitigate this problem. Keywords: initial public offering, anti-takeover provisions, firm value, closely held shares
1 Statement of Originality
This document is written by Student Chantal Raap who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents
2 Table of Contents
1. Introduction ... 3
2. Literature Review ... 5
2.1 Definitions and reasoning ... 5
2.2 Theoretical insights ... 6
2.3 Empirical evidence ... 8
2.4 Conclusion and hypotheses ... 9
3. Methodology and Data ... 10
3.1 Methodology ... 10
3.2. Data and descriptive statistics ... 14
4. Analysis ... 19
4.1. Empirical Results ... 19
4.2 Robustness check ... 23
5. Conclusion and discussion ... 25
3 1. Introduction
Over the last number of years, corporations and countries can be characterized as becoming more globalized and internationalized. This resulted in greater opportunities for
corporations, but also increased the importance of a good corporate governance structure. Fortunately, the current literature on corporate governance has also established over the years. They have studied diverse elements of countries’ financial systems, but after decades of research the effects of anti-takeover mechanisms remain quite uninvestigated.
This study will contribute to the literature on corporate governance and anti-takeover provisions. More specifically, this paper will focus on the effect of closely held shares on firm’s value at initial public offering date. Closely held shares are the shares within a publicly traded company held by shareholders who are directly affiliated with the
company or management. Owners of closely held shares participate in management or control of the firm (Feld 1974). Closely held shares cannot be publicly traded. This implies that companies who have a higher amount of closely held shares are more resistant to hostile takeovers, since the shares are held within a small, interested group of shareholders.
This paper provides a closer look on the effect of closely held shares at the initial public offering date (IPO) on a firm’s value, using data from firms that went public between 1999 and 2015. Thus, the hypothesis that will be tested is: is there an impact of closely held shares on IPO firm value in the Netherlands, Belgium and United Kingdom in the years 1999-2015? The overalleffect will be tested, as well as the effect in the individual countries. The initial public offering date of a firm is an interesting time period to examine the effect of anti-takeover mechanisms, since this is the stage in which firms are most likely to use anti- takeover provisions. Moreover, at the firm's first listing date, a firm is not yet limited by their previous choices of implemented takeover mechanisms.
The academic objective of this research is to create new insights in the literature of anti-takeover provisions and the effect on firm’s value at initial public offering date. This study contributes to current research in multiple ways. First of all, available research on the impact on IPO firm value is mostly outdated, especially for European countries. For instance, the latest study from the Netherlands examined a time period from 1984 till 1999. This study will focus on a substantially different time period from 1999 to 2014. Furthermore,
4
not much recent research has been done on the comparison between countries. In general, the literature only compares European data to data from the US. Since countries have different legal and corporate governance structures, comparing different western European countries could give interesting new insights. This study is not only interesting in an
academic way, but also in the sense of practical relevance. Shareholders and top
management could learn from the impact of closely held shares on IPO firm’s value. It can inform owners on the costs and benefits of closely held shares. They can use this research to make better decisions in the distribution of the shares. Moreover, if managers act on behalf of the shareholders there’s less need for a tighter corporate governance structure and shareholders should interfere less with management. On the other hand, if managers entrench themselves, a tight corporate governance structure needs to be implemented to take account for managerial entrenchment.
There exists sufficient empirical evidence on the question whether public firms adopt takeover defenses. Literature suggests that more than 90% of public firms adopt at least one takeover provision before going public (Roosenboom and van der Groot 2003). The unanswered question remains whether there is a positive or negative effect of these provisions, since literature provides contradictory results. The existing debate on the impact of takeover defenses on IPO firm value is one of the most controversial aspects of corporate governance. Therefore, it is rather interesting to investigate whether closely held shares have an impact on IPO firm’s value.
This paper will provide a closer look on the effect of closely held shares at the initial public offering date of a firm using data from firms that went public between 1999 and 2015. The final data set consist of 203 observations, including data from the Netherlands, Belgium and the United Kingdom. Outliers have been removed and industries combined. All financial data, as well as information on the amount of closely held shares, are collected from Datastream. Using this data a few regression analyses will provide an in-depth view of the impact of closely held shares. These regressions will also give information on the
differences within the three countries. The combination of the three countries gives no significant results, though the sign is negative. The results on the regressions separating the countries show that in the United Kingdom there is a significant effect of closely held shares in IPO firm value. The robustness check creates new insights on the lack of significance of the combined countries. It appears that this is due to the credit crunch and Internet bubble.
5
The proceedings of the paper are as following. First, the main theories and empirical findings in existing literature comparable to this study will be examined. This section will give a prediction on the outcome of this study as well as a theoretical and empirical
background. The section will also contain the main hypothesis and sub-questions. Secondly, the methodology will describe the model and data collection method. After this, the
empirical findings will be exposed and discussed. This section is followed by a robustness check, which provides new insights. At last, the conclusions, limitations and advice for further research will be made.
2. Literature Review
This section will provide a theoretical and empirical overview of prior literature. First, it will define anti-takeover mechanisms, closely held shares and reasoning for examination at IPO stage. Moreover, the corporate governance structure within the Netherlands, Belgium and United Kingdom will briefly be explained. After this, the two main theoretical theories on takeover mechanism and corporate governance will be studied. Also, the stockholder interest theory and the entrenchment theory will be examined. Moreover, the empirical findings of previous literature will be provided and finally the hypotheses will be
determined.
2.1 Definitions and reasoning
Anti-takeover mechanisms are implemented with the intention to prevent the firm from a hostile takeover. A hostile takeover occurs when investors go to the company’s
shareholders to get the acquisition approved, instead of coming to an agreement with the company’s management. When the investor obtains more than 50% of the shares, he can replace management and take all control of the company.
There are various types of takeover defenses distinguished. Majority of types of takeover defenses have been studied in prior literature. Though, there is still a lot to explore on this topic. Prior literature provides evidence that closely held shares can be seen as an anti-takeover mechanism (Daines and Klausner 2001; Roosenboom and van der Groot 2003). Therefore, this study will provide new insights on closely held shares as an
anti-6
takeover mechanism.
Reasoning why this study investigates the effect at IPO stage is motivated in the previous section. These reasons are also considered in prior literature. Coates (2000) studied a sample of 320 IPO’s in 1991-1992 and he concludes that anti-takeover provisions are optimal at their IPO stage. Furthermore, managers want to ensure that their benefits will be protected while the firm goes public. Therefore, they are likely to use anti-takeover provisions to ensure their benefits (Brennan and Franks 1997). Dorsman and Gounopoulos (2013) provide evidence on the impact of financial crises on Dutch IPO’s. They find that IPO’s are underpriced in Europe during the credit crunch and Internet bubble, though existing literature on anti-takeover provisions at IPO stage is considered more essential.
This study will take a closer look at the effect on IPO firm value of the combination of the Netherlands, United Kingdom and Belgium. Moreover, the paper also provides
comparison between the Netherlands, Belgium and United Kingdom. Since there are different law and corporate governance structures, this could create interesting new
insights. Previous studies have investigated the differences between countries and come up with four types of corporate governance structures including the Anglo-Saxon system, the Germanic system and the Latin system (De Jong 1991; Weimer and Pape 1999; Moerland 1995). The Netherlands is classified as the Germanic system, the United Kingdom as the Anglo-Saxon system and Belgium as the Latin system. Weimer and Pape (1999) describe that shareholders of the Anglo-Saxon countries have most influence on managerial decision making. Their article provides an international framework for corporate governance structures. They find that shareholders in Latin and Germanic countries have less influence on managerial decision making, though in Latin countries they are more influential than in Germanic countries.
2.2 Theoretical insights
The focus of this study is on the effect of anti-takeover mechanisms. Much research has been done on the problem of corporate governance, though results are inconclusive. The findings of prior literature rely on some theories which support some general hypotheses. The first theory is the principal-agent theory. Grossman and Hart (1980) studied whether managers act in the interest of the owners. They describe the problem as following. The
7
owner of a firm is the principal and the agent is the manager. The primary assumption is that the owner cannot monitor the manager’s action. However, the owner can observe the outcome of the manager’s actions. The outcome is considered the firm’s profit, which does not only depend on the manager’s actions, but also on other factors outside the manager’s control. Thus, if the firm does well, the owner will not have a clear view on whether this is because of the manager or because the manager has been lucky. It is difficult to monitor the manager to ensure they act on the principal’s behalf. Muth and Donaldson (1998)
empirically test the validity of the principal agency theory and stewardship theory of listed Australian companies in 1994. They state that the issue of the agency problem is that it became more complex for shareholders to sustain overview over the corporation.
Moreover, shareholders lack necessary inside knowledge which managers are aware of. This results in managers pursuing their personal goals, instead goals in favor of the corporation. If a hostile takeover occurs, the management will be replaced. So the threat of a takeover reduces the manager’s interest in fulfilling company goals instead of personal goals. Although the principal-agency theory seems to be the most dominant theory
underlying governance research, there are theoretical limits to the theory (Hirsch, Michaels, & Friedman 1987). According to Davis, Schoorman and Donaldson (1997), the stewardship theory defines that managers are stewards whose motives are aligned with the company owners, instead of their personal goals. This implies that managers prefer to act to the interest of the company. This theory doesn’t assume that the stewards have no personal objectives. Instead of the agency theory, the stewardship theory states that managers value the objectives of the company over their personal objectives. Problems between
management and owners can only occur when corporate structures are not well defined (Davis et al. 1997). The problem of shareholder lacking inside knowledge, as Muth and Donaldson (1998) describe, does not exist since shareholder will now benefit from managers having inside knowledge. This theory states that managers only implement anti-takeover provisions if these provisions benefit the shareholders, given that there are no corporate structure problems.
In order to understand the importance of takeover defenses, the discussion concerning takeover mechanisms needs to be explained. There exist two traditional
hypotheses that argument this conflict. The managerial entrenchment hypothesis refers to the principal-agency theory. According to DeAngelo and Rice (1983), this hypothesis states
8
that managers will protect themselves from a hostile takeover at the cost of the
shareholders. Reflecting this view, managers will use anti-takeover mechanisms without regard to shareholders, resulting in a lower firm value (Johnson et al. 2015).
Contrary to this view is the shareholder interest hypothesis. For instance, Stulz (1988) empirically analyses the impact of managerial voting rights on firm value. He states that anti-takeover mechanisms will benefit shareholders and therefore increase firm value. The shareholder interest hypothesis assumes that the shareholders gain from a higher price while having more takeover defenses implemented (DeAngelo and Rice, 1983). This is because the managers will have a greater bargaining position due to a potential bidding war between acquires, according to Jarrel et al. (1988).
2.3 Empirical evidence
Field and Karpoff (2002) examine the motives and consequences of the implementation of anti-takeover provisions at IPO stage. They use a sample of more than 1000 IPO’s in the US and regress on possible determinants of anti-takeover provisions at the IPO stage. They conclude that managers of an IPO firm adopt takeover defenses at the time of their IPO. They mention that managers with a greater amount of shares are more likely to act in the benefit of the shareholders, since a loss in firm value would also be at personal costs of the manager. Furthermore, Jensen and Meckling (1976) demonstrate who bears the costs of anti-takeover provisions. They find evidence in line with Field and Karpoff (2002). They also provide same reasoning and state that the agency problem will be mitigated and thus increase the firm value.
Contrary to previous conclusions, a more recent study concludes different effect. Souther (2015) uses a sample of closed-end funds and examines the impact of takeover defenses on IPO firm value. Closed-end funds use the same defenses as general
corporations, but compete in a more homogeneous environment. His sample included a 15-year period consisting of 622 US funds. Souther (2015) reports that anti-takeover provisions decreases firm value. This implies that managers extract benefits from shareholders.
Managers use the anti-takeover provisions to protect their personal benefits in terms of compensation levels. The study finds that compensation levels are higher while more takeover defenses are used. These findings are in line with the managerial entrenchment hypothesis.
9
Another study of Keloharju and Kulp (1996) uses a sample of Finnish initial public offerings to discover the effect of shares owned by management. They test the agency hypothesis by examining the relationship between management ownership and market-to-book value at IPO stage. Their sample consists of 60 IPO’s in the period of 1960-1993. They find that management ownership and IPO firm value don’t have a singular positive or negative effect. Keloharju and Kulp (1996) conclude that shares owned by management are significantly positively related to IPO firm value at low percentage of shares owned by management. Though, when ownership by management is high they don’t find significant effect, although the sign is positive. This implies some evidence for the agency theory, but only at low levels of management ownership.
Keloharju and Kulp are not the only study concluding both a positive and negative effect of closely held shares on IPO firm value. Their results are also supported from literature in the United States. For instance, McConnell and Servaes (1990) empirically analyze two samples of more than 1000 firms and find a significant curvilinear relationship of management ownership and firm value. If the amount of closely held shares is below 40%, there exists a positive relationship on firm value. However, if it is above 40%, a negative relationship appears.
2.4 Conclusion and hypotheses
Prior studies don’t show a singular answer on the effect of anti-takeover provisions on IPO firm value. The principal-agent theory and the stewardship theory create two views on the behavior of managers to shareholders benefits. Two general hypotheses on the impact of takeover defenses are derived from this. The managerial entrenchment hypothesis states that anti-takeover mechanisms are at the cost of shareholders and therefore decrease firm value. On the other hand, the shareholder interest hypothesis states that managers will have a better bargaining position while implementing anti-takeover provisions, such that these provisions will increase firm value. Previous literature has shown that there exist evidence for both the managerial entrenchment hypothesis and the shareholder interest hypothesis. This study will also contribute to this debate by testing the impact of shares held by management. Depending on which hypothesis will be supported, shareholders will be able to derive the optimal company structure to optimize firm value.
10
Again, the main hypothesis of study will answer the following question: is there an impact of closely held shares on IPO firm value in the Netherlands, Belgium and United Kingdom in the years 1999-2015? Moreover, if empirical evidence finds that there is impact, is this a positive of negative impact? This will determine whether this study will provide evidence for the managerial entrenchment hypothesis or the shareholder interest hypothesis. The sub-questions of this study involve the country specific effects. Like explained before, the three countries differ in their corporate governance structure which makes it interesting to investigate if the impact differs across these western European countries. Therefore, the sub-questions will be: is there an impact of closely held shares on IPO firm value in the Netherlands in the years 1999-2015? Is there an impact of closely held shares on IPO firm value in the United Kingdom in the years 1999-2015? Is there an impact of closely held shares on IPO firm value in Belgium in the years 1999-2015? Naturally, the outcome will be framed within the theoretical background.
3. Methodology and Data
This section describes the methodology and sample data. This contains a description of what type of data is selected, which variables will be used and which observations are excluded with economic reasoning. Moreover, the regression method and econometric model will be included together with the statistical hypothesis. It will examine the causal effect of de dependent and main independent variable used and gives predictions about the variables used. Also it will provide a definition of the interpretation of significance and the sign in relation with answering the hypothesis. Furthermore, it will examine the data source and it will describe what will be adjusted and/or removed in terms of the data. Also, this section will provide the summary statistics for the main hypothesis and sub-questions. Finally, it will also contain a cross-correlation table between the dependent and independent variables and a VIF to test for multicollinearity.
3.1 Methodology
The first regression and set of data will be based on the main hypothesis. The research question that this study aims to answer, like established in the literature review, is: Is there an impact of closely held shares on IPO firm value in the Netherlands, Belgium and United
11
Kingdom in the years 1999-2015? The corresponding sub questions have the same
regression method; though differ by using different data sets. When combining the answers on the three sub-questions, this study will be able to expose whether there are differences on the impact between these three western European countries.
In terms of data this study demands different kinds of data from around the time period of the initial public offering date of a firm. The first necessary data to find will be all the companies in the Netherlands, United Kingdom and Belgium that went public for the first time. After having all available companies from the database, companies outside the time frame of 1999-2015 as an initial public offering date should be removed from the data set. The next step is to use the combination of companies and IPO dates to find all first-day closing prices, as well as the number of shares outstanding (at the day of IPO). Finally, the total assets, net sales or revenues, total debt, EBITDA and closely held shares need to be generated from the database. This presents all yearly data of every firm from 1999-2015. The total assets in the financial year before IPO have to be selected, as well as the total sales in the year before and year of IPO, total debt in the year before IPO, EBITDA in the year before IPO and the number of closely held shares in the year of IPO.
In order to be able to answer the hypothesis, an econometric model has to be defined. This study will implement an ordinary least square regression. This is an extensively used statistical method widely used in empirical finance to estimate unknown parameters in a linear regression model (Vogelsang and Wagner, 2014). OLS minimizes the differences between the observed variables and predict the linear approximation of the variables. The econometric model that will be used in this research is described as following:
𝐿𝑛 (𝑉𝐴𝐿𝑈𝐸𝑖) = 𝛽0+ 𝛽1 𝐿𝑛 (𝑇𝐴𝑖) + 𝛽2𝐿𝐸𝑉𝑖 + 𝛽3𝑆𝐺𝑖 + 𝛽4𝑅𝑂𝑆𝑖 + 𝛽5𝑀𝐴𝑁𝑈𝑖 + 𝛽6𝑆𝐸𝑅𝑉𝑖 + 𝛽7𝐼𝑁𝐷𝑖 + 𝛽8𝐶𝐻𝑆𝐻𝐴𝑅𝐸𝑆𝑖+ 𝜀𝑖
This economic model consists of a dependent variable, seven control variables and the percentage of closely held shares. The dependent variable is the natural logarithm of firm value. This is computed using the market capitalization, calculated as the first-day closing price times the number of shares outstanding immediately after IPO. Next, the market capitalization is divided by EBITDA. This measurement of calculating firm value is conforming to the measurement used by Smart et al. (2003). To adjust for size effects, the natural
12
logarithm of firm value will be calculated.
The first independent control variable is total assets in the year before IPO.
Comparable to firm value, this will also be calculated by taking the natural logarithm of total assets. The second control variable is leverage. Leverage will be computed by dividing the total debt with the total assets, both in the year before IPO. This study expects that leverage will increase the firm’s value. If a manager attracts more (long-term) debt, the free cash flows of the manager will decrease. This implies that the manager cannot spend the free cash flow on harmful investments to a firm’s value. Following, the percentage of sales growth will be computed using the total sales in the year of IPO minus the total sales in year before IPO. This variable is included to control for the possibility that a firm with higher growth perspective is valued higher. This amount will be divided by the total sales in the year before IPO. Return on sales is also added to take account for probability differences. This is calculated by the EBITDA divided by total sales, in the year before IPO. The last three control variables are dummies on the firm’s industry. The three industries are
manufacturing, service and industrial industries. If a firm’s divided in one of the three industries, the dummy equals 1 and if not, the dummy equals zero. All these independent variables are in line with studies from Johnson, Karpoff and Yi (2015), Smart et al. (2003) and Roosenboom and van der Groot (2003) on anti-takeover mechanisms on firm value. The percentage of closely held shares is the main variable of interest. This will be computed by dividing the amount of closely held shares with the amount of shares outstanding at the IPO.
In order to know to which extend a causal relationship exists between de percentage of closely held shares and the value of a firm, prior literature has to be examined. Johnson et al. (2015) state that takeover provisions are not only correlated with a higher firm value, but also a cause of a higher firm value. They used instrumental variables and find that the relationship of a firm’s value at the IPO stage and the use of anti-takeover mechanisms have a positive relation. This research differs in the fact that some of the control variables are different, as well as the anti-takeover provision used. However, since the shares outstanding and number of closely held shares are measured at the same time, it can be assumed that there is causal relationship to some extent.
13
statistical hypothesis, which can be used for both my main hypothesis and sub-questions, is described as following:
𝐻0: 𝛽8 = 0 𝐻1: 𝛽8 ≠ 0
The 𝐻0 implies that there will be no effect of closely held shares on IPO’s firm value and the 𝐻1 implies that there is effect of closely held shares. Assuming that the percentage of closely held shares are significant the following holds: if 𝛽8 will be greater than zero, there will be a positive effect on firm value. This supports the shareholder interest hypothesis. Entire the same, if 𝛽8 will be smaller than zero there will be a negative effect on firm value, supporting the managerial entrenchment hypothesis. Furthermore, the interpretation of the coefficient and the effect on firm value is described by the following formula:
𝑅𝑎𝑡𝑖𝑜 𝐹𝑖𝑟𝑚 𝑉𝑎𝑙𝑢𝑒 = 𝐹𝑖𝑟𝑚 𝑉𝑎𝑙𝑢𝑒𝑛𝑒𝑤
𝐹𝑖𝑟𝑚 𝑉𝑎𝑙𝑢𝑒𝑜𝑙𝑑 = 𝑒
𝛽8∗Δ𝐶𝑙𝑜𝑠𝑒𝑙𝑦 ℎ𝑒𝑙𝑑 𝑠ℎ𝑎𝑟𝑒𝑠
For significance, a t-test will be conducted to identify whether and on what extend the 𝛽8 explains the relationship between the closely held shares and firm value. If the effect of 𝛽8 turns out not to be significant, no conclusions can be made upon the effect of closely held shares. This is because no statistically significant linear dependence can be observed. There could be no linear relationship between the two variables, or the regression method needs to be adjusted.
This study expects that the relation between closely held shares and firm value is positive, as reported by with Jensen and Meckling (1976) and Field and Karpoff (2002). Moreover, studies from Keloharju and Kulp (1996) and McConnell and Servaes (1990) find that if the percentage of closely held shares is low, there exists a positive effect on firm value. Since the sample shows that the percentage of closely held shares is quite low, this study assumes that firm value will rise. As Jensen and Meckling (1976) mention in their studies, if shareholders who are directly affiliated with the company have a higher amount of shares, they also bear for the cost of moral hazard. The more shares they own, the higher the alignment of incentives of managers and (unrelated) shareholders. The agency problem will be migrated and this expectation is in line with the shareholder interest hypothesis. This
14
study predicts that the effect will be the same within the western European countries. Though, since the countries have different corporate governance structures, this study expects that the effect will be largest in in the Netherlands, followed by Belgium and the United Kingdom. This would be in line with prior literature described in the previous section. 3.2. Data and descriptive statistics
All financial data on IPO companies are collected from Datastream. Datastream is a financial database which contains data on financial business, stock prices and macroeconomics. The database contains more than 175 countries with a history of up to 50 years back. It contains all listed companies for the years 1999-2015 required for this paper. The Dutch IPO
company data is collected from Euronext Amsterdam. Similarly, data from the United Kingdom is collected from Financial Times Stock exchange and data from Belgium is collected from Euronext Brussels.
The original (combined) sample contained 365 observations. Some observations are excluded from the data set, since they display different characteristics compared to other IPO firms. For example, this includes lots of companies from financial or utilities sectors. This is caused by different regulations within these industries. Others were excluded due to the lack of information availability. For instance, the amount of closely held shares wasn’t available for every observation in the database. This variable couldn’t be found in any other databases, so it had to be excluded. After these exclusions, the sample contained 208 observations.
All data from the Netherlands and Belgium was published in Euro’s, but since data from the United Kingdom was published in Pounds, this had to be alternated to Euro’s. Therefore, the historical exchange rates per day had to be exhaled from the European Central Bank (2016). Using these daily exchange rates, all the observations from the United Kingdom could be substituted to Euro’s.
The sample included numerous different industries. With regard to being able to take account for industrial effects, this study combined the smallest industry groups to the major industry groups. In this manner, dummies could be included to make the regression more valid. For instance, following from the combination of consumer services,
telecommunication and health care, the services industry became one of the dummies. Concerning having the most reliable results as possible, the data set containing 208
15
observations had to be shortened with some outliers. Outliers increase standard errors and therefore reduce the reliability of the results. They are illegitimately included and therefore should be removed (Anscombe, 1960). Outliers are found through Stata, by creating
boxplots of all the variables and looking at z-statistics. After eliminating outliers, which had an impact considering the standard errors and significance and sign of the regression analysis, the final data set contained 202 observations. Table 1 will expose the summary statistic for the whole sample, table 2 will summarize statistics for the Netherlands, table 3 for the United Kingdom and table 4 summarizes statistics for Belgium.
Furthermore, table 5 will provide a cross-correlation table of the dependent and independent variables to test for multicollinearity. The numbers are Pearson correlation coefficients. Their range is from minus one till one, whereas the closer to one, the higher the correlation. A negative value indicated an inverse relationship to the correlated value. There’s no reason to assume that there would be multicollinearity, since correlations between variables are low and one of the industry dummies will always be excluded from the regression.
Nevertheless, there exists criticism on this method of checking for multicollinearity since the matrix doesn’t take account for indirect correlation. Moreover, there is no unilateral indication on when there exists a multicollinearity problem. Therefore,
multicollinearity is also tested by using a variance inflation factor. The advantage of using this method is that the VIF is considered reliable. It is a widely used method to check for multicollinearity in OLS regressions. A VIF test measures how much the variance of the regression is increased because of multicollinearity. If VIF is greater than 4 or 1/VIF is smaller than 0.25, there could be a multicollinearity problem. Table 6 provides the results on the VIF-test on the independent variables. Concluding from this table, it is assumable that there exist no multicollinearity problems.
Finally, graph 1 will expose the number of IPO’s per year. The horizontal axis defines the years and the vertical axis defines the number of IPO’s. The amount of IPO’s are
separated in the specific countries, whereas the blue column represents the number of IPO’s in the Netherlands. The red column represents the IPO’s in the United Kingdom and the green column represents the number of IPO’s in Belgium. The graph shows that there seems to be a lower amount of IPO’s during times of economic crises.
16
Table 1: Summary statistics, including mean, standard deviation, minimum and maximum for the combined observation (n=202).
Note: data on the summary statistics from this table is collected from Datastream.
Table 2: Summary statistics, including mean, median, standard deviation, minimum and maximum for observations from the Netherlands (n=28).
Note: data on the summary statistics from this table is collected from Datastream.
Combined N Mean Median Standard
Deviation
Minimum Maximum
Market capitalization/EBITDA (€ x1000)
202 887.17 539.30 2,759.78 -20,772.72 12,123.84
Closely held shares (%) 202 0.04 0.03 0.06 0.00 0.42 Total Assets (€ x1000) 202 6,190.43 419.94 24,000 2.99 203,000
Leverage (%) 202 34.92 31.35 32.25 0.00 184.47
Sales growth (%) 202 64.86 13.01 307.65 -73.09 3,298.61 Return on sales (%) 202 12.62 16.52 69.50 -551.78 245.42 Price-to-sales ratio 202 1.30 0.05 13.98 0.00 198.37
Netherlands N Mean Median Standard
Deviation
Minimum Maximum
Market capitalization/EBITDA (€ x1000)
28 18.29 12.09 89.40 -272.51 325.63
Closely held shares (%) 28 0.03 0.03 0.03 0.00 0.11 Total Assets (€ x1000) 28 8,437.58 262.95 293.00 2.99 145,000.00
Leverage (%) 28 26.50 21.79 26.10 0.00 83.87
Sales growth (%) 28 103.55 15.68 251.44 -59.01 969.77 Return on sales (%) 28 19.66 14.44 65.71 -173.72 245.42
17
Table 3: Summary statistics, including mean, median, standard deviation, minimum and maximum for observations from the United Kingdom (n=139).
Note: data on the summary statistics from this table is collected from Datastream.
Table 4: Summary statistics, including mean, median, standard deviation, minimum and maximum for observations from Belgium (n=35).
Note: data on the summary statistics from this table is collected from Datastream.
United Kingdom N Mean Median Standard
Deviation
Minimum Maximum
Market capitalization/EBITDA (€ x1000)
139 1,283.51 978.82 3,253.34 -20,772.72 12,123.84
Closely held shares (%) 139 0.04 0.03 0.06 0.00 0.42 Total Assets (€ x1000) 139 6,261.73 544.00 25,200.00 6.02 203,000.00
Leverage (%) 139 36.71 29.95 35.31 0.00 184.47
Sales growth (%) 139 64.95 13.32 351.55 -73.09 3,298.61 Return on sales (%) 139 16.04 16.98 59.72 -551.78 133.64 Price-to-sales ratio 139 0.40 0.06 1.37 0.00 11.27
Belgium N Mean Median Standard
Deviation
Minimum Maximum
Market capitalization/EBITDA (€ x1000)
35 8.23 7.00 26.86 -29.35 142.40
Closely held shares (%) 35 0.05 0.03 0.07 0.00 0.31 Total Assets (€ x1000) 35 4,111.33 131.51 12,000.00 9.08 62,400.00
Leverage (%) 35 34.59 38.39 21.86 0.00 77.13
Sales growth (%) 35 33.58 8.45 73.94 -45.43 350.38 Return on sales (%) 35 -6.61 15.51 100.93 -312.25 148.40 Price-to-sales ratio 35 5.80 0.01 33.51 0.00 198.37
18
Table 5: Cross correlation table of the dependent and all independent variables of the regression analysis.
Note: ** and * indicate significance at 5% and 1%, respectively.
Table 6: VIF test on the independent variables to check for multicollinearity.
VALUE TA SG ROS MANU SERV IND CHSHARES
VALUE 1.0000 TA 0.2326** 0.0009 1.0000 SG -0.2729** 0.0001 -0.0685 0.3327 1.0000 ROS 0.5303** 0.0000 0.1517* 0.0312 -0.0800 0.2575 1.0000 MANU 0.0168 0.8121 0.1481* 0.0354 -0.0120 0.8651 0.0813 0.2502 1.0000 SERV -0.0677 0.3576 -0.1107 0.1168 -0.0105 0.8820 -0.0662 0.3489 -0.5561** 0.0000 1.0000 IND 0.0651 0.3576 0.0022 0.9749 0.228 0.7479 0.0077 0.9132 -0.2098* 0.0027 -0.6959** 0.0000 1.0000 CHSHARES -0.0550 0.4369 0.0131 0.8530 -0.0118 0.8672 0.0415 0.5572 0.1479* 0.0357 -0.1351 0.0552 0.0312 0.6594 1.0000 VIF 1/VIF Closely held shares 1.03 0.98 Total Assets 1.05 0.95 Leverage 1.05 0.95 Sales growth 1.03 0.98 Return on sales 1.06 0.95 Service 2.05 0.49 Industrial 1.98 0.51 Mean VIF 1.32
19 0 5 10 15 20 25 Netherlands United Kingdom Belgium
Graph 1: The number of IPO’s per year, separated between the Netherlands, United Kingdom and Belgium.
4. Analysis
This section will test the main hypothesis and sub-questions as described in the literature review. The hypotheses will be resolved using four regression analyses computed from Stata. First, some conditions have to be verified. Explanatory tables will be given and values interpreted combined with economic reasoning. Moreover, results will be placed in more general context and implications will be discussed. After this, the next paragraph will do a robustness check in order to assess the reliability of the results along with reasoning for the non-existence of significance. In essence, a robustness check has already been done using regression analyses of the countries separated instead of the combined sample as a whole. These country specific regressions had some comparable results, but some variables showed contradictorily output. Therefore an alternative robustness check will be exposed.
20 4.1. Empirical Results
As mentioned in the previous paragraph, outliers have been removed beneficial to the significance of the variables. By looking at the boxplots of all different variables, massive outliers were straightforward to identify and were removed. Moreover, to ensure that no multicollinearity problems exist between the dummies for different industries, one of them is omitted in every regression. In terms of the other variables, the correlation matrix along with the VIF test has proven that no multicollinearity problems exist. The optimal omitted industry for the total sample, the sample for the Netherlands and United Kingdom, is the manufacturing industry. For Belgium the industrial industry is the omitted variable. This study assumed heteroscedasticity. To control for this, the robust standard errors are computed in the regression analysis.
To answer the hypotheses of this study, four regression analyses had to be
performed. The outputs of these regressions are converted into table 7. Table 7 shows the results of four ordinary least square regressions, using the logarithm of firm value as a dependent variable and the percentage of closely held shares as the main independent variable. The regressions include 7 control variables, though one of the industry dummies is omitted due to multicollinearity. The variables in parentheses represent the t-statistics, using the robust standard errors.
21
Table 7: regression analyses of the combined countries as well as the separated countries.
Note: If a variable is significant at the 10% level the t-values are followed by *; if significant at the 5% level, values are followed by **; if significant at 1% level, values are followed by ***.
Inconsistent with the expectations of this research, we find that the impact of closely held shares have no significant results in the combined countries, since the tvalue equals -1.25. Moreover, this also accounts for the sample from the Netherlands and Belgium (t-values 1.63 and 1.55, respectively). These results suggest the non-existence of any impact on IPO firm value. Apparently, closely held shares do not yield incentive effects for these regressions.
The absence of significant results is surprising. The lack of significance could be due to the authorization of closely held shares. The explanation could be that managers will only possess shares if there is a threat of a hostile takeover. This suggests that this anti-takeover provision will only be implemented at times with existence of a hostile takeover threat, instead of at the IPO date. On the other hand, the absence of significant results for the
Firm Value
Combined Netherlands United Kingdom Belgium Closely held shares
-5.77 (-1.25) -9.32 (1.63) -5.64 (-1.92)* -12.75 (1.55) Control variables Total assets 0.29 (2.20)** 0.46 (1.63) -0.11 (-1.07) 0.43 (2.25)** Leverage 0.26 (0.35) -3.37 (-0.80) -0.22 (-0.42) -0.76 (-0.45) Sales growth -0.32 (-8.90)*** -0.50 (0.27) -0.37 (4.87)*** 0.87 (1.91)* Return on sales 3.07 (6.04)*** 4.08 (4.40)*** 2.84 (4.87)*** 2.62 (7.76)*** Manufacturing -0.33 (-0.42) Service 0.29 (0.42) -1.96 (-0.95) 0.62 (0.74) -1.20 (-1.45) Industrial 0.96 (1.26) -1.89 (-0.98) 1.06 (1.31) Intercept 7.17 (3.49)*** 4.40 (0.83) 14.09 (6.52)*** 3.63 (1.55) R² 0.37 0.37 0.52 0.75 Root MSE 3.52 3.85 2.30 2.13 F-value 27.87 8.84 122.24 35.57 N 202 28 139 35
22
Netherlands and Belgium could partially be on behalf of the sample size. The sample for the Netherlands and Belgium is relatively small, which increases the standard errors and thus reduces the probability of significance of closely held shares. Yet, this doesn’t provide reasoning for the combined sample size, which includes 202 observations. Furthermore, the R² are 0.37, 0.37, 0.52 and 0.75 respectively. This implies that the models are reliable, since the threshold of at least a R² of 0.2 is reached conveniently. This interpretation for instance of the combined sample, is that 37% of the variance in firm value can be explained by the independent variables.
Nevertheless, there is a significant relation on the effect of closely held shares in the United Kingdom at the 10% significance level (t-value of -1.92). The impact of closely held shares is negatively related to the IPO firm’s value, which implies that the percentage closely held shares reduce the firm value at IPO stage. This identifies that the stewardship theory doesn’t hold, while the agency problem extends. Moreover, the managerial entrenchment hypothesis gets support by these results.
The negative sign of the coefficient accounts for all countries, so the percentage closely held shares seems to be inversely related to IPO firm value. This consolidates that the effect of closely held shares on IPO firm’s value differs across countries, which provides even more support for the managerial entrenchment hypothesis. This is yet another
surprising result, since the effect was expected to be positive. Maybe this contrary outcome is due to the years studied in this paper. Prior literature which supported the expected outcome used different time periods, while literature using a more recent time period is more in line with this outcome.
The differences in impact of the separated regression analyses are partly surprising. The regressions show that closely held shares have most influence in Belgium, followed by the Netherlands and United Kingdom. Weimer and Pape (1999) found that Anglo-Saxon countries had most influence on managerial decision making. This implies that closely held shares would have the least impact. This is in line with the findings of this study. On the other hand, the expectations for the Netherlands and Belgium differ from prior literature.
For the control variables total assets, sales growth and return on sales, the effects are significant in the first regression (t-values of 2.20, -8.90 and 6.04, respectively). This is in line with the described literature. Remarkably, the United Kingdom lacks significance on total assets. Moreover, within the Netherlands total assets and sales growth are not all
23
significant; this could be due to the small sample size. Leverage is not significant for any of the regressions. The sign of the country specific regressions is contrary compared to the expectation of study. This study assumed that the impact of reduced free cash flows would increase on firm’s value. This only seems to hold while looking at the combined sample of the three countries.
Surprisingly, the dummies accounting for industrial effect do not have significant effect. This is contrary to other literature, whereas industries do have an effect. This could be related to the distribution of the various industries among the three dummies. Further research should try different combinations of industries to ensure the absence of
significance.
In conclusion, the outcome of the combined regression examines that there exist no effect of the percentage of closely held shares on IPO firm’s value. Most of the
sub-questions conclude the same effect, except for the United Kingdom. The United Kingdom provides significant evidence to state that there exist a negative effect of the percentage of closely held share on IPO firm value. The sign of the combined countries as well as the specific countries is inversely related to IPO firm’s value in every regression. The following paragraph will examine a robustness check to examine the correctness of these findings. 4.2 Robustness check
The robustness check in this section will create new insights. Previous literature states that economical changes occurred during the time period used in this paper. More detailed, Dorsman and Gounopoulos (2013) describe in their article that there is impact on IPO firm value of the Credit crisis and Internet bubble. He reports that IPO firms are underpriced in times of low economies. This also affected the western European countries examined in this study. Due to the possibility of consequences of these crises on the results of this study, the robustness check will inspect whether results would differ among these events. The same time frame is applicable.
The combined sample will be divided within two different time periods. One
regression will examine the combined period of the dot com bubble and credit crisis within the time frame of 1999-2015. The other will examine the period without any crisis. The dot com bubble includes the years 1999-2001 for the combined sample and for the credit crisis,
24
the regression includes observations from 2002-2009. The credit crisis affected these western European countries since December 2009. Since there are no observations within December 2009, the time frame will take account for all the observations from 2010-2015. The regression output of the ordinary least square is summarized in table 8. Table 8 shows results of the combined countries as a whole, divided into two time frames. These regressions also used the logarithm of firm value as a dependent variable and the
percentage as closely held shares as the main independent variable. The regression still includes 7 control variables. Within the first two time period, the industrial dummy is omitted due to multicollinearity. The third time period excludes the manufacturing dummy to take account for multicollinearity. The variables in parentheses represent the t-statistics, again using robust standard errors.
Table 8: the regression analyses of the combined sample using different time periods as robustness check.
Note: If a variable is significant at the 10% level the t-values are followed by *; if significant at the 5% level, values are followed by **; if significant at 1% level, values are followed by ***.
Combined Firm Value
2002-2009 1999-2001 2010-2015 Closely held shares -15.02
(-2.05)** -2.42 (-0.48) Control variables Total assets -0.04 (-0.24) 0.51 (2.53)** Leverage 0.32 (0.33) 0.43 (0.36) Sales growth -0.29 (-6.98)*** -0.30 (-6.86)*** Return on sales 3.85 (5.86)*** 2.75 (4.88)*** Manufacturing 0.92 (0.99) -2.15 (-2.01)** Service -1.27 (-1.28) -0.08 (-0.12) Industrial Intercept 12.80 (4.17)*** 4.85 (1.84)* R² 0.46 0.38 Root MSE 3.31 3.58 F-value 20.40 23.73 N 94 108
25
The results of these regressions are consistent with prior literature and support the managerial entrenchment hypothesis. When looking at the time period whereas no crises exist, the percentage of closely held shares is significant (t-value of -2.05). This implies that there exists an inverse relationship with IPO firm value. Contrary to the time period 2002-2009, the period of credit crisis and internet bubble don’t show significant results.
Reasoning could be that managers do not expect hostile takeover threats during periods of crisis. Therefore, they don’t need to implement anti-takeover provisions. But when there is a normal economy, the chance of a takeover increases relatively to times of economic depression. Managers will take account for possible takeover threats by attracting more shares, which will decrease the chances of a takeover. These findings are in line with previous suggestion that managers will only implement anti-takeover provisions at time with higher takeover threats.
5. Conclusion and discussion
This paper focused on the effect of closely held shares, seen as an anti-takeover provision, on firm value at IPO stage. This effect was studied by using a final sample consisting of 202 observations, using data from the Netherlands, United Kingdom and Belgium. The main hypothesis was to test the effect on the combined countries. After this, three regressions on the country specific effects were tested. The result of the combined countries was not significant, though the sign was negative. Moreover, this also applied to the Netherlands and Belgium. For the United Kingdom there exists significant effect.
Prior literature is based on two theories: the principal-agent theory and the stewardship theory. From these theories, two hypotheses derived: the managerial entrenchment hypothesis and the shareholder interest hypothesis. There exists literature that provides evidence for both theories. This study, however, provides evidence for the managerial entrenchment theory. This is not as expected, since this study expected that firm value would increase. The main reasoning for this was that when managers attract more shares, they would also expect firm value to rise since they want to gain from their investment. This was consistent with some prior literature. However, prior literature consistent with the expectation of this study turned out to be outdated, since the results were only consistent with more recent studies.
26
Furthermore, this study is partially in line with the findings of Weimer and Pape (1999). Their article describes that shareholders of the Anglo-Saxon countries have most influence on managerial decision making. This implies that due to this influence, the impact on firm value should be the lowest. This is in line with the analysis described in section four. On the other hands, the analysis is not in line with the expectations on the Netherlands and Belgium.
This paper didn’t provide significant results of closely held shares for the combined analysis. This lack of significance could be explained by the suggestion that managers will only implement anti-takeover provisions at times with higher chances of a hostile takeover, instead of IPO date. This suggestion is supported by the robustness check. This section provided significant results of the impact of closely held shares on firm value in the combined analysis. This implies that in times without economic depression there is a negative effect on IPO firm value, while in times of economic depressions, no conclusions can be based due to the lack of significance.
This study provides practical relevance for shareholders and managers. Shareholders and managers should be aware that the amount of closely held shares will decrease firm value. Therefore shareholders should find a method to keep the amount of closely held shares on a low level. Like examined before, this study supports the managerial
entrenchment hypothesis. Managers will not act on the behalf of shareholders and
therefore lower firm value to obtain their personal goals. This implies that there is need for a tight corporate governance structure.
Limitations of this study are that the sample size had to be narrowed down due to information availability. Since information of closely held shares was only available in one database, some observations had to be excluded while they could have increased the significance of the results. This is a selection bias and further research could try to hand-collect these data. Furthermore, industry dummies don’t have a significant effect. This was contrary to prior literature and could be due to the combination of industries combined to the three dummy variables. This could also be improved in further research.
27 Reference list
Anscombe, F. J., 1960, Rejection of outliers, Technometrics 2(2), 123-146.
Brennan, M. J., & Franks, J., 1997, Underpricing, ownership and control in initial public offerings of equity securities in the UK, Journal of Financial Economics 45(3), 391- 413.
Coates, I. V., & John, C., 2000, Explaining variation in takeover defenses: Failure in the corporate law market, Harvard Law and Economics Discussion Paper, 297.
Daines, R., & Klausner, M., 2001, Do IPO charters maximize firm value? Antitakeover protection in IPOs, Journal of Law, Economics, and Organization 17(1), 83-120.
Davis, J.H., Schoorman, F.D., Donaldson L., 1997, Toward a stewardship theory of Management, Academy of Management review 22-1, 20-47.
DeAngelo, H., Rice, E. M., 1983, Antitakeover Charter Amendments and Stockholder wealth, Journal of Financial Economics, 329-359.
De Jong, H. W., 1991, The takeover market in Europe: Control structures and the
performance of large companies compared, Review of Industrial Organization 6(1), 1-18.
Dorsman, A., & Gounopoulos, D., 2013, European sovereign debt crisis and the performance of Dutch IPOs, International Review of Financial Analysis 30, 308-319.
European Central Bank, 2016, Euro exchange rates GBP, January 10th 2016, on
https://www.ecb.europa.eu/stats/exchange/eurofxref/html/eurofxref-graph- gbp.en.html
28
Feld, A. L., 1974, The Implications of Minority Interest and Stock Restrictions in Valuing Closely-Held Shares, University of Pennsylvania Law Review, 934-953.
Field, L. C., & Karpoff, J. M., 2002, Takeover defenses of IPO firms, Journal of Finance, 1857-1889.
Grossman, S. J., & Hart, O. D., 1980, Takeover bids, the free-rider problem, and the theory of the corporation, The Bell Journal of Economics, 42-64.
Hirsch, P., Michaels, S., & Friedman, R., 1987, "Dirty hands" versus "clean models", Theory
and Society 16, 317-336.
Jarrell, G. A., Brickley, J. A., & Netter, J. M., 1988, The market for corporate control: The empirical evidence since 1980, The Journal of Economic Perspectives, 49-68.
Jensen, M. C.,1988, Takeovers: Their Causes and Consequences, The Journal of
Economic Perspectives 2(1), 21-48.
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.
Johnson, W. C., Karpoff, J. M., Yi, S., 2015,. The bonding hypothesis of takeover defenses: Evidence from IPO firms, Journal of Financial Economics 117, 307-332.
Keloharju, M., & Kulp, K., 1996, Market-to-book ratios, equity retention, and management ownership in Finnish initial public offerings, Journal of Banking & Finance 20(9), 1583-1599.
McConnell, J. J., & Servaes, H., 1990, Additional evidence on equity ownership and corporate value. Journal of Financial economics 27(2), 595-612.
29
Moerland, P. W., 1995, Corporate ownership and control structures: An international comparison, Review of industrial organization 10(4), 443-464
Muth, M., & Donaldson, L., 1998, Stewardship theory and board structure: A contingency approach, Corporate Governance: An International Review 6(1), 5-28.
Roosenboom, P., Groot, T., 2003, Takeover defences and IPO Firm Value in the Netherlands,
European Financial Management 9(4), 485.
Smart, S. B., Thirumalai, R. S., & Zutter, C. J., 2008, What's in a vote? The short-and long-run impact of dual-class equity on IPO firm values, Journal of Accounting and Economics 45(1), 94-115.
Souther, M.E., in press, The effects of takeover defenses: Evidence from closed-end funds, Journal of Financial Economics.
Stulz, R., 1988, Managerial control of voting rights: Financing policies and the market for corporate control, Journal of financial Economics 20, 25-54.
Taylor, S., & Whittred, G., 1998, Security design and the allocation of voting rights: Evidence from the Australian IPO market, Journal of Corporate Finance 4(2), 107-131.
Vogelsang, T. J., & Wagner, M., 2014, Integrated modified OLS estimation and fixed- inference for cointegrating regressions, Journal of Econometrics 178(2), 741-760.
Weimer, J., & Pape, J., 1999, A taxonomy of systems of corporate governance, Corporate
