MASTER THESIS: Ownership concentration, institutional ownership, and innovation. An empirical study.

(1)

MASTER THESIS:

Ownership concentration, institutional ownership, and

innovation. An empirical study.

By Marc Mohr

Field Key Words: Corporate Governance, Innovation, Institutional Ownership, Ownership Concentration, Shareholders’ rights

Study Program: MSc International Financial Management Faculty of Economics and Business

(2)

2

Abstract

This thesis both investigates the effect of ownership concentration on innovation, as well as the effect of institutional ownership on innovation by analyzing a sample consisting of 7,952 firms and 39,738 firm-year observations in 48 countries from all continents. Further, this study examines the effect that firms operating in high technological sectors have on both relationships and how the level of shareholders’ rights in a country affect the aforementioned relationship. Combining agency and resource dependency theories, we argue that institutional ownership positively affects the level of innovation in a firm and that this relationship is strengthened when a firm operates in a high technological sector. Meanwhile, there is no statistically significant evidence indicating an effect of ownership concentration on the level of innovation of a firm. And lastly, while the level of shareholders’ rights in a country do impact innovation directly, the results show no moderating effect on either of the relationships.

(3)

3

Abstract

2 1. Introduction

4 2. Literature review and hypothesis development

6 2.1 Ownership concentration and innovation

7 2.2 Institutional ownership and innovation

8 2.3 High technological sector

10 2.4 Shareholders’ rights

11 3. Methodology

12 3.1 Data and sample

12 3.2 Methodology and variables

13

(4)

4

1. Introduction

The importance of innovation for firms is widely studied in the literature. The first studies surrounding innovation are from the early 19th_{century (Architect, 1802) and nowadays, it is still a hot topic. Studies}

about innovation differ severely from the main drivers (Del Canto, and Suarez, 1999; King et al., 1994; Rothwell, 1990) to the result or consequences (Alina et al., 2003; Merton 1992). All are contributing to the literature in their way.

But what is innovation exactly, and what is the importance of innovation? There is not one answer to the first question as economists have many, different definitions for innovation. According to O’Sullivan, and Dooley (2009) “Innovation is the process of making changes, large and small, radical and incremental, to products, processes, and services that result in the introduction of something new for the organization that adds value to customers and contributes to the knowledge store of the organization.” Thus, via innovation, not only firms but also customers can increase their wealth. The necessity of innovation is also acknowledged by governments, as the Department of Trade commented on the link between continuous innovation and jobs, profit and standard of living: ‘’If UK-based companies fail to innovate, jobs and profits will suffer, and our standard of living will fall compared with other countries’’.

(5)

5 This thesis will focus on ownership concentration and institutional ownership. To be more precise, this thesis will answer the following research question:

“What effect does ownership concentration has on the level of innovation?” “What effect does institutional ownership has on the level of innovation?”

Former studies predominantly focus on one country or market but this thesis uses a sample of 48 countries in all continents, thereby contributing to the existing literature on ownership structures and innovation. Moreover, the sample period is from 2002 until 2013 with nearly 40,000 firm-year observations. Additionally, this thesis investigates the moderating effect of firms operating in a high vs. low technical sector and what the effect is of shareholders’ rights on the relationship between ownership concentration and innovation, and institutional ownership and innovation.

(6)

6

2. Literature review and hypothesis development

The main theoretical framework for explaining the relationship between ownership and innovation has been the agency theory (Francis and Smith, 1995; Choi et al., 2011) and little attention has been paid to the resource dependence theory. This thesis will combine these theories in one framework to further, and more explicit, explain the relationship between ownership and innovation. More precise, the relationship between ownership concentration and innovation and the relationship between institutional ownership and innovation will be examined in this thesis.

Jensen and Meckling (1976) describe the agency relationship as ‘’a contract under which one or more persons (the principle) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent’’. In this thesis, the principles are the owners of the company, and the agents are the management of that company and agency costs arise if the principle and agent do not act in the same manner. Holmström (1982) argues that there are agency costs related to innovation because innovations are risky, unpredictable, long-term, labor-intensive and idiosyncratic, causing a possible discrepancy between the owners and the managers of a company on whether or not to pursue certain investments. The riskiness is due to the high probability of failure and the chances of exceptional returns. The unpredictability comes from the many on-foreseeable features during innovation processes. Also, innovations are usually long-term processes and often exist of multiple stages. These different stages consist of human capital making innovation labor intensive and, lastly, innovations are idiosyncratic as they are often firm or project specific and not easily comparable with, or duplicated by others.

(7)

7 overcome this influence. The first part of the RDT describes the necessity of certain resources for any organization to operate successfully and that its environment will affect that organization in a manner of ways. These resources are not only materialistic or financial but can also be laws and regulations or a stable politic climate, i.e., all factors that influence an organization’s way to success. The second part of the RDT focuses on the attempt of an organization to become independent and to get rid of the uncertainty of its environment (Hillman et al., 2009). As described by Holmström (1982), innovation itself has a lot of characteristics that are not easily met. For certain organizations, innovation is near impossible due to lack of technology, finance, or laws and regulations (Klein and Sorra, 1996). A particular ownership structure could liberate an organization from its dependence on its existing environment by offering a new set of availabilities for the organization. In other words, resource-rich outside shareholders bring in the missing pieces of the puzzle needed for innovation to the firm.

2.1 Ownership concentration and Innovation

Previous literature has shown a positive relationship between concentrated ownership and innovation (Deng et al., 2013). Large shareholders have a positive effect on strategic management as it increases the efficiency in that management, can more easily monitor the actions of that management and can also influence the actions of that management (Francis and Smith, 1995). This power to control and influence management that large shareholders have increases an organization’s performance and results in an incentive to increase the innovation (Choi et al., 2012). Also, the more concentrated large shareholders there are, the less friction between owners there is, and in this way, a set of unified objectives can be given to the management. This contributes to a reduction in agency costs and therefore leads to an increase in innovation.

(8)

8 and Stole (1993) argue that minority stakeholders are short-term driven and therefore only impelled by the price of the shares. Next to this, they are sensitive to any take-over gain and also not interested in long-term, risky R&D investments. But large shareholders, on the other hand, are more driven by information (Lee and O’Neill, 2003) and therefore understand long-term R&D investments and goals from the management. They are also not likely to sell their bulk of shares and therefore provide stability to an organization (Choi et al., 2012).

Thus, concentrated ownership provides a unified objective to managers and it leads to a reduction of agency costs. Also, large shareholders are not short-term driven and provide stability to an organization. Therefore, we hypothesize:

H1: The level of concentrated ownership has a positive effect on innovation

2.2 Institutional ownership and innovation

(9)

9 Another theory is the lazy manager theory. This theory claims that managers are ‘’lazy’’ and lack incentives to maximize profits, as they do not actively search for new ways to enhance value for their firm. Hart (1983) acknowledges this by arguing that managers prefer a quiet life and that institutional investors force them to innovate. McConnel and Sarvaes (1990) back this argument by showing that institutional owners influence management, both financial and via other mechanisms.

Organizations are dependent on specific resources, which ultimately originate from their close environment (Pfeffer, 1972). Innovations require particular resources that firms often do not possess, but that are in the hands of other organizations (Choi et al., 2012). To come in possession of these resources, firms have to link with these other organizations. In this light, institutional ownership provides a useful solution in a firm’s quest for innovation-related resources. Institutions can provide finance, technology, and knowledge and so help the firm to increase its innovation activities (Chen et

al. 2013).

Lastly, Eng and Shackell (2001) argue that institutional owners are long-term orientated and therefore support firms that engage in innovation activities (Amsden, 1989; Kim, 1997; Wade, 1990). Hall and Lerner (2009) show in their study that institutional ownership improves asymmetric information problems connected with research and development of a firm. Both are contributing to a positive relationship between institutional ownership and innovation.

Institutional ownership overthrows the fear of managers of being fired after bad performance from innovation due to tight control and monitoring, and institutional owners force ‘’lazy’’ managers into innovation. Also, as institutional owners are long-term orientated, and they can provide the missing resources needed for a firm to innovate, the third hypothesis is:

(10)

10

2.3 High technical sector

Firms operating in a sector that is considered as a high technological sector are home to higher rates of innovation (Thornhill, 2006; Hsu et al., 2014). This is due to the higher level of intangible assets, a more skilled workforce, and a higher level of creativity (Kwon, and Yin, 2006). Also, these firms are involved in businesses where innovation is more present such as design, implementation, and the release of new products due to the high technical nature of these sectors (Hsu et al., 2014).

This thesis will use firms operating in a high technological sector as a firm-level moderator for the relationships between ownership concentration and innovation, and institutional ownership and innovation. More precise, this thesis predicts that the level of innovation will be higher when a firm is active in a high technological sector. Thus, it will have a strengthening effect on the relationship between ownership concentration and innovation, as the large shareholders know the market that they operate in, is more technological and, thus, innovation is key if they want to compete. When more ownership lies with only a few equity holders, the easier decisions are made, and the higher the innovation will be in such a technological sector. Therefore the third hypothesis is:

H3a: The positive effect of ownership concentration is greater for firms operating in high technological sectors.

For institutional ownership, the same logic holds. Knowing that the firm operates in a high technological sector, institutions will be prepared to act on this by increasing the level of R&D expenses. Otherwise, it would have invested in another sector. Therefore we hypothesize:

H3b: The positive effect of institutional ownership is greater for firms operating in high

(11)

11

2.4 Shareholders’ rights

Countries with high shareholders’ rights provide a certain level of security and a stable environment for investors. This feeling of security will lead to more liquidity in the market (Levine, 1997). According to Alfaro et al. (2004) and La Porta et al. (1997), shareholders’ rights and creditor rights are the basis for financial development. A developed financial market will lead to a reduction of moral hazard, adverse selection and transaction costs, causing a more optimal allocation of capital (Rajan, and Zingales, 1998), which, in turn, results in lower costs for firms to attract capital for investment projects, which will consequently increase the level of innovation.

Large listed firms often do not have equal voting powers amongst their shareholders, as large shareholders and shareholder’s pacts dominate the shareholder meetings (Lhuillery, 2011). However, better shareholders’ rights make sure that the democratic practices are followed more strictly and that minority shareholders will be able to execute their voting power. This increase in the power of minority shareholders leads to higher agency costs as now more “principles” need to agree on how the firm is managed — leading to lower innovation activities.

Thus, shareholders’ rights can both strengthen and weaken the relationship between institutional ownership and innovation. It can increase innovations as investors feel more secure in lending their money, resulting in lower costs for firms borrowing capital. However, with more shareholder’ rights, the agency costs will rise, leading to lower innovation.

As shareholders’ rights increase the power of minority shareholders, the power of majority shareholders decreases. Weakening the strength that a few large shareholders have, will lead to less efficient decision making and lower innovation. Especially, the positive effect of ownership concentration on innovation will be weakened by shareholders’ rights. Thus, the fourth hypothesis is:

(12)

12 For institutional ownership, the development of shareholders’ rights is an ambiguous one. Investors feel safer, and the costs of attracting capital for investments decreases for firms, leading to more R&D expenditures and, eventually, more innovation. But the power increase of minority shareholders leads to less efficient decision making and lower innovation. Therefore the we hypothesize the following:

H4b: The level of shareholders’ rights has a strengthening effect on the relationship between institutional ownership and innovation

H4c: The level of shareholders’ rights has a weakening effect on the relationship between institutional ownership and innovation

3. Methodology

3.1 Data and sample:

(13)

13

3.2 Methodology and variables:

To test the hypotheses, we estimate ordinary least square (OLS) regressions. Furthermore, country, year, and industry fixed effects are incorporated in the regression, as well as robust standard errors that are clustered at the country level. In the next section, all the variables and regressions will be explained.

3.2.1 Dependent variable:

The dependent variable in this thesis is innovation. Innovation can be measured in terms of input as the level of R&D (Hall and Bagchi-Sen, 2002), in terms of output via the number of patents held by a firm (Seifert, and Gonenc (2012); (Francis et al. (2011); (Chang et al. (2015)) or a by looking at the R&D intensity, patents divided by R&D expenses (Baysinger, and Hoskisson, 2017). Due to a lack of data on patents of firms outside the United States, this thesis will focus on innovation in terms of input, measured as the level of R&D expenses of a firm divided by its total assets. Also, we were not able to figure out the length of R&D projects for each firm in the sample, so the level of R&D is purely an accounting measure. Furthermore, when firm-year observations were unknown, we did not exclude them from the sample, but rather followed Francis et al. (2011) by giving them the value of zero.

3.2.2 Independent variables:

(14)

14

3.2.3 Firm-level moderator:

This thesis will use firms that operate in a high technology sector as a firm-level moderator. Following Choi et al. (2012), we construct a dummy variable that takes the value of 1 if firms operate in the chemical, automotive, microelectronics, pharmaceutical, and communications industries, and zero otherwise. Firms in the high technology sector are usually associated with higher levels of R&D expenditures and innovative projects.

3.2.4 Country-level moderator:

The country-level moderator on the relationship between ownership concentration and innovation and the relationship between institutional ownership and innovation in this thesis is the level of shareholders’ rights within a country. Following Demirgüç-Kunt and Maksimovic (2002), we use the shareholder rights index, spanning from 0 to 10. Zero meaning no shareholders’ rights in a country and ten meaning extremely well-established shareholders’ rights. Following La Porta et al. (1998), in a given country in a given year, the shareholders’ rights score is given to all firms within that country.

3.2.5 Control-variables:

(15)

15 with different currencies. A profitable firm is more able to engage in innovative activities as it can reinvest the returns it makes (Kochhar, and David, 1996). The return on assets is used as a proxy for profitability and calculated as the net profit divided by total assets, and the sales growth is calculated as the annual growth rate of sales (Choi et al., 2012; Lee, and O’Neill, 2003). The market-to-book ratio is used as a proxy for investment opportunities (Lee, and O’Neill, 2003), and a higher market-to-book ratio should lead to higher R&D expenses. Also, as debt reduces a firm’s appetite for long-term investments, higher leverage should indicate lower R&D expenses (Jensen and Showalter, 2004). Therefore, leverage is also used as a control variable in this thesis. Lastly, the country-level variable GDP per capita in US dollars is used as a control variable.

_𝑡

+ 𝜇

_𝑖

+ 𝜀

_𝑖,𝑡

This thesis focuses on coefficient

β

1

in model 1, which measures the sensitivity of R&D to change

in OWNCON, to test hypothesis 1. A positive and significant coefficient means that an increase

in OWNCON leads to a higher R&D. For hypotheses 2, the focus is again on model 1 and the

coefficient β

1

. However, the variable OWNCON will be substituted for INSTITOWN. Now a

1_{For regression model 1) and 2) the independent variable OWNCON is substituted for INSTITOWN to test}

(16)

16

positive and significant coefficient shows the increase of R&D to a positive change in

INSTITOWN.

For hypothesis 3a and 3b, we will also focus on model 1. Coefficient β

9

will show the

moderating effect of HTS on the relationship between OWNCON and R&D. To accept

hypothesis 3a, coefficient β

9

is expected to be positive as to show the strengthening effect

that HTS has on the relationship between OWNCON and R&D. For hypothesis 3b, we will look

at coefficient β

9

again, but now to look at the effect of HTS on the relationship between

INSTITOWN and R&D. Again, a positive coefficient is expected and in line with the hypothesis.

Lastly, to test hypotheses 4a, 4b, and 4c, model 2 will be used. To test hypothesis 4a, the focus

will be on coefficient β

11

. According to the hypothesis, the coefficient is expected to be

negative, to show the weakening effect of SHARE on the relationship between OWNCON and

R&D. For hypothesis 4b and 4c, the focus will be on coefficient β

11

. For hypothesis 4b (4c), the

estimate is expected to be positive (negative), as to show the strengthening (weakening)

effect of SHARE on the relationship between INSTITOWN and R&D.

4. Results

4.1 Descriptive statistics

(17)

17

In table 2, the descriptive statistics for the dependent, independent, and country-level

variables are shown per country. The limits of this dataset are visible as Pakistan, Kuwait,

United Arabian Emirates, and Morocco have less than ten firm-year observations. This has to

do with the lower availability of data on ownership in these countries. Further, the United

States, Japan and the United Kingdom cover over half the sample with 15,013, 3,838 and 3,048

firm-year observations, respectively. Sixteen countries show a 0.000 for the mean of R&D, and

the sample mean is 0.010. Denmark has the highest R&D level with a mean of 0.062, followed

by Switzerland and the United States with values of 0.038 and 0.034, respectively.

Looking at the OWNCON, we see that the United States, Canada, and the United Kingdom

have low means. This is in line with the fact that commonwealth countries are often associated

with high diversified ownership. Ownership in these countries is primarily built on stocks and

not on banks or other institutions. Pakistan and Morocco show high figures for OWNCON.

However, both countries only have data on one firm and therefore no real conclusion can be

drawn from this. The Philippines, Russia, and the Czech Republic all show a mean of nearly 70

percent. Compared to an overall mean of 45.1. Looking at INSTITOWN, we see that Turkey

Table 1. Descriptive distribution by year

YEAR N Percentage 2002 2,664 6.70 2003 2,743 6.90 2004 2,914 7.33 2005 3,112 7.83 2006 3,225 8.12 2007 3,259 8.20 2008 3,422 8.61 2009 3,655 9.20 2010 3,767 9.48 2011 3,847 9.68 2012 3,847 9.68 2013 3,283 8.26 Total 39,738 100.00

(18)

18

Table 2. Descriptive statistics by country

R&D OWNCON INSTITOWN GDP SHARE

Country N Mean N Mean N Mean N Mean N Mean

(19)

19

shows the lowest mean of 0.3 and Ireland and the United States have the highest with 10.1

and 10.8 respectively. Seven countries show a score of 0.00, and the overall mean is 3.44

percent.

The country-level variable GDP does not show surprising figures. Pakistan and India have the

lowest GDP with a score of 7.06 and 7.09, and United Arabian Emirates, Norway and

Switzerland are in the top with a GDP of over eleven. The sample mean is 9.75. For the variable

SHARE, which is an index between 0-10, we see that the Philippines are the only country that

scores a zero. The Republic of Korea scores a three and the United States and Portugal both

score a 4. Canada, Russia, and Spain all have a score of 9 and Hungary has the highest score

of 10.

Table 3 – Panel A shows the descriptive statistics for the whole sample. Table 3 – Panel B

shows the number of firm-year observations and means for the firm-level regression variables

separated for firms operating in low- and high technological sectors. In total, the sample has

31,254 (79%) firm-year observations operating in low technological sectors and 8,484 (21%)

in high technological sectors. To overcome any disruptions caused by the, strongly present,

outliers, this thesis follows Seinert, and Gonenc (2012) by winsorizing the firm-level variables

at the top and bottom 1%.

The dependent variable R&D has an average of 0.02 with a standard deviation and median of

0.05 and 0.00. The average is significantly higher in high technological sectors than in the low

ones (0.07 vs. 0.008). This result is in the line of expectations as firms performing in the high

technological sector are expected to engage more in R&D activities (Hsu et al., 2014).

(20)

20

mean (28.32) and the firms in the high technological sector score lower (24.347). INSTITOWN

spreads from zero until 96 percent, with a mean of 6.73. Firms operating in low technological

sectors show a mean slightly below the sample average of 6.319 and the firms operating in

the high technological sector show a mean 8.254.

It is noticeable that the other variable means of firms operating in low- and high technological

sectors are always different at a one percent significance level. The ROA is higher for firms

operating in low technological sectors. This is not in line with the theoretical background, as

Kochhar, and David (1996) argued that more profitable firms, have a higher ROA, are more able to

Table 3 – Panel A. Descriptive statistics

N Mean St. Dev. Min Median Max

R&D 39,738 0.02 0.05 0.00 0.00 0.64 OWNCON 39,738 28.32 24.48 0.00 22.87 100.00 INSTITOWN 39,738 6.73 9.42 0.00 0.00 96.00 ROA 39,738 0.06 0.10 -0.55 0.06 0.37 SALESGRO 39,738 0.16 0.42 -0.49 0.08 3.58 ASSETS 39,738 14.68 1.62 9.18 14.69 18.53 MTB 39,738 1.38 1.40 0.09 0.93 8.09 LEV 39,738 0.31 0.22 0.00 0.32 0.90 HTS 39,738 0.18 0.38 0.00 0.00 1.00 GDP 39,738 10.36 0.83 6.97 10.64 11.19 SHARE 39,738 5.59 0.73 0.00 3.00 4.00

Note: This table provides the number of firm-year observations, mean, standard deviation, minimum, median and maximum for the dependent variable R&D; independent variables OWNCON and INSTITOWN; firm-level variable HTS; country-level variable SHARE; and control variables ROA, SALESGRO, ASSETS, MTB, LEV and GDP. Definitions and sources for the variables are given in the Methodology section.

low technological sector high technological sector Mean Difference

N Mean N Mean R&D 31,254 0.008 8,484 0.070 0.062*** OWNCON 31,254 29.392 8,484 24.347 5.045*** INSTITOWN 31,254 6.319 8,484 8.254 1.935*** ROA 31,254 0.065 8,484 0.051 0.014*** SALESGRO 31,254 0.153 8,484 0.180 0.027*** ASSETS 31,254 14.810 8,484 14.221 0.589*** MTB 31,254 1.211 8,484 1.984 0.733*** LEV 31,254 0.340 8,484 0.219 0.121***

Note: This table provides the count and mean of the firm-level regression variables, split into firms that operate in the low- and high technological sector. The mean difference is calculated with a univariate test, where *, **, *** represent a statistical significance level of 10, 5 or 1 percent.

(21)

21 reinvest those profits in innovative projects and here, the opposite holds. Firms operating in high technological sectors do show a higher SALESGRO and MTB. The higher MTB ratio is in line with the theory, as Lee, and O’Neill (2003) argue that MTB can be seen as a proxy for investment opportunities and a higher MTB ratio should indicate higher R&D expenses. At last, firms operating in low-technological sectors show a higher mean of LEV. The higher level of LEV is in line with the literature as a higher debt reduces a firm’s appetite for long-term investment (Jensen and Showalter, 2004).

4.2 Correlation analysis.

Table 4 shows the correlation matrix for this sample. The independent variable OWNCON is negatively correlated with R&D. This is not according to previous literature and also not in line with the hypotheses. INSTITOWN, however, is positively correlated with R&D and therefore, in line with the hypothesis and previous literature. The firm-level variables ROA, ASSETS, and LEV correlate negatively with R&D, while SALESGRO, MTB, and HTS correlate positively. Both the country-level variables correlate positively with the dependent variable R&D. The high correlation between OWNCON and INSTITOWN does not pose a problem as they are not in the same regression, so there is no risk of collinearity.

Table 4. Selected sample Correlation

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) 1) R&D 1.000 2) OWNCON -0.1192* 1.000 3) INSTITOWN 0.1430* -0.2012* 1.000 4) ROA -0.2720* 0.0848* -0.0581* 1.000 5) SALESGRO 0.0371* 0.0609* -0.0107 0.0395* 1.000 6) ASSETS -0.2147* -0.0735* -0.1601* 0.0566* -0.1599* 1.000 7) MTB 0.2692* 0.0277* 0.0682* 0.2432* 0.1695* -0.4173* 1.000 8) LEV -0.2233* -0.0264* -0.0600* -0.1151* -0.0688* 0.4009* -0.4128* 1.000 9) GDP 0.1439* -0.4384* 0.1744* -0.1600* -0.0671* 0.0244* -0.0448* -0.0141* 1.000 10) HTS 0.4995* -0.0845* 0.0842* -0.0579* 0.0255* -0.1486* 0.2265* -0.2217* 0.0693* 1.000 11) SHARE -0.1673* 0.3298* -0.2598* 0.1287* 0.0064 0.0834* -0.0407* 0.0393* -0.4865* -0.1065* 1.000

(22)

22 4.3 Multivariate analysis

Table 5 shows the results of regressing the level of ownership concentration on innovation as well as the moderating effect of firms operating in high technological sectors, while controlling for firm- and country-level characteristics. Model 2 shows a significant and negative relationship between OWNCON and R&D, which is not in line with the previous literature and therefore contradicting Hypothesis 1. Moreover, in Model 3, the coefficient of OWNCON is no longer significant, indicating that we cannot say with certainty if the relationship is positive or negative. This means that we reject Hypothesis 1. Model 3 further shows that the coefficient on HTS*OWNCON is significant and negative, meaning that the effect of OWNCON on R&D is lower for firms operating in high technological sectors. Again, this is

Table 5. Ownership concentration and innovation

Model 1 R&D Model 2 R&D Model 3 R&D OWNCON -0.00002** -0.00002** -0.00007 (0.000) (0.000) (0.000) ROA -0.02282*** -0.02282*** -0.02296*** (0.004) (0.004) (0.004) SALESGRO 0.00043 0.00043 0.00045 (0.000) (0.000) (0.000) ASSETS -0.00258* -0.00258* -0.00261* (0.001) (0.001) (0.001) MTB 0.00144*** 0.00144*** 0.00145*** (0.000) (0.000) (0.000) LEV -0.00487*** -0.00487*** -0.00488*** (0.001) (0.001) (0.001) GDP 0.00452*** 0.00452*** 0.00460*** (0.001) (0.001) (0.001) HTS -0.02093*** -0.01870*** (0.006) (0.006) HTS * OWNCON -0.00009*** (0.000) Constant 0.00590 0.00590 0.00582 (0.016) (0.016) (0.016)

Country FE Yes Yes Yes

Industry FE Yes Yes Yes

Year FE Yes Yes Yes

Adjusted R2 0.4753 0.4753 0.4772

Observations 33,761 33,761 33,761

(23)

23 not in line with recent literature, and we, therefore, must reject Hypothesis 3a. Lastly, the coefficient on HTS is significant and negative, meaning that firms operating in a high technological sector have a negative impact on R&D.

Table 6 shows the results of regressing the level of institutional ownership on innovation as well as the moderating effect of firms operating in high technological sectors, while controlling for firm- and country-level characteristics. Model 4 and 5 show a positive and significant coefficient on INSTITOWN, meaning that the level of institutional ownership positively affects innovation. However, in Model 6, with the introduction of the interaction term HTS*INSTITOWN, the coefficient on INSTITOWN is no longer significant. Model 6 does show a positive and significant (at 1% level) coefficient on HTS*INSTITOWN, meaning that firms operating in a high technological sector positively influence the

Table 6. Institutional ownership and innovation

Model 4 Model 5 Model 6

INSTITOWN 0.00008*** 0.00008*** 0.00002 (0.000) (0.000) (0.000) ROA -0.02267*** -0.02267*** -0.02294*** (0.004) (0.004) (0.004) SALESGRO 0.00045 0.00045 0.00044 (0.000) (0.000) (0.000) ASSETS -0.00253* -0.00253* -0.00253* (0.001) (0.001) (0.001) MTB 0.00145*** 0.00145*** 0.00148*** (0.000) (0.000) (0.000) LEV -0.00498*** -0.00498*** -0.00496*** (0.001) (0.001) (0.001) GDP 0.00461*** 0.00461*** 0.00466*** (0.001) (0.001) (0.001) HTS -0.02108*** -0.02381*** (0.006) (0.006) HTS * INSTITOWN 0.00026*** (0.000) Constant 0.00293 0.00293 0.00257 (0.015) (0.015) (0.015)

Year FE Yes Yes Yes

Adjusted R2 0.4761 0.4761 0.4801

Observations 33,761 33,761 33,761

Note: this table reports the estimated coefficients from regression firm-level variables, and firm-level moderators, on R&D. Definitions of variables are provided in the methodology. Models employ country, year and industry fixed effects, these are not shown in the regression, but can be asked for. The regressions use robust standard errors corrected for clustering at the country-level. Sample period is 2002-2013. Standard errors are shown in the brackets. *,**,*** stand for the statistical significance levels of 10%, 5% and 1% respectively.

(24)

24 relationship between institutional ownership and innovation. This is in line with previous literature as investors of institutions know that they need to innovate if they want to keep up in these sectors (Hsu

et al., 2014), causing for higher R&D expenses. The positive and significant coefficient on

HTS*INSTITOWN also confirms Hypothesis 3b.

Table 7 reports the results of the moderating effect of shareholders’ rights on the relationship between ownership concentration and innovation while controlling for firm- and country-level characteristics. First, Model 9 shows a negative and insignificant coefficient on OWNCON, meaning that we must,

Table 7. Ownership concentration and innovation

OWNCON -0.00007 -0.00004 -0.00005 (0.000) (0.000) (0.000) ROA -0.02296*** -0.02278*** -0.02277*** (0.004) (0.004) (0.004) SALESGRO 0.00045 0.00041 0.00041 (0.000) (0.000) (0.000) ASSETS -0.00261* -0.00247* -0.00247* (0.001) (0.001) (0.001) MTB 0.00145*** 0.00146*** 0.00146*** (0.000) (0.000) (0.000) LEV -0.00488*** -0.00493*** -0.00493*** (0.001) (0.001) (0.001) GDP 0.00460*** 0.00319*** 0.00317*** (0.001) (0.001) (0.001) HTS -0.01870*** -0.01562** -0.01552** (0.006) (0.006) (0.006) HTS * OWNCON -0.00009*** -0.00008*** -0.00009*** (0.000) (0.000) (0.000) SHARE -0.00250*** -0.00243*** (0.000) (0.000) SHARE * OWNCON -0.00001 (0.000) Constant 0.00582 0.02881 0.02859 (0.016) (0.020) (0.020)

Year FE Yes Yes Yes

Adjusted R2 0.4772 0.4797 0.4796

Observations 33,761 33,761 33,761

(25)

25 again, reject Hypothesis 1. The interaction variable HTS*OWNCON remains negative and significant, leading us to reject Hypothesis 3a. In Model 8, the variable SHARE is introduced to the model, increasing the explanatory power of the model with 0.2%. The coefficient on SHARE is significant and negative, meaning that it harms innovation. The coefficient on SHARE*OWNCON is negative and insignificant. Therefore we cannot say if there is a positive, negative, or no effect of shareholders’ rights on the relationship between ownership concentration and innovation. Thus, rejecting Hypothesis 4a.

(26)

26 Overall, the results show no significant effect of ownership concentration on innovation, leading us to reject Hypothesis 1. The effect of institutional ownership on innovation is insignificant in Model 6, but significant and positive in Model 12. The adjusted R2_{is higher in Model 12 than in Model 11 and Model}

6, indicating a higher explanatory power of the Model. The positive coefficient of INSTITOWN in Model 12 indicates a positive effect on R&D. This is in line with recent literature and Hypothesis 2. Accordingly, Aghion et al. (2013) argue that managers can choose between low-risk short-term investments and riskier long-term investments. Recent literature shows that managers will choose low-risk short-term investments due to their focus on quarterly results (Porter, 1992) and their fear of being fired after negative results from long-term investments (Kaplan, and Minton, 2006). However, Choi et al. (2012)

Table 8. Institutional ownership and innovation

INSTITOWN 0.00002 0.00001 0.00012* (0.000) (0.000) (0.000) ROA -0.02294*** -0.02277*** -0.02275*** (0.004) (0.004) (0.004) SALESGRO 0.00044 0.00039 0.00040 (0.000) (0.000) (0.000) ASSETS -0.00253* -0.00241* -0.00240* (0.001) (0.001) (0.001) MTB 0.00148*** 0.00149*** 0.00149*** (0.000) (0.000) (0.000) LEV -0.00496*** -0.00500*** -0.00500*** (0.001) (0.001) (0.001) GDP 0.00466*** 0.00325*** 0.00327*** (0.001) (0.001) (0.001) HTS -0.02381*** -0.02071*** -0.02047*** (0.006) (0.006) (0.006) HTS * INSTITOWN 0.00026*** 0.00025*** 0.00023*** (0.000) (0.000) (0.000) SHARE -0.00245*** -0.00234*** (0.000) (0.000) SHARE * INSTITOWN -0.00002 (0.000) Constant 0.00257 0.00268 0.02584 (0.015) (0.019) (0.019)

Year FE Yes Yes Yes

Adjusted R2 0.4801 0.4824 0.4825

Observations 33,761 33,761 33,761

(27)

27 argue that these arguments do not hold when institutional owners come into play, as they are more active in monitoring and controlling managers and therefore reduce the uncertainty of managers that are willing to partake in innovative activities, leading to an increase in innovation.

The results from Model 3 and 9 show a negative and significant coefficient for HTS*OWNCON, meaning that firms operating in a high technological sector have a decreasing effect on the relationship between ownership concentration and innovation. This is not in line with recent literature, as they indicated a positive effect. Therefore we must reject Hypothesis 3a. The results in Model 6 and 12 show a positive and significant coefficient of HTS*INSTITOWN, meaning that the positive effect of INSTITOWN on R&D is strengthened when firms operate in high technological sectors. This is in line with recent literature and we therefore accept Hypothesis 3b.

Lastly, this sample does not provide evidence for a moderating role of shareholders’ rights on the relationship between the level of ownership concentration and innovation, or the relationship between the level of institutional ownership and innovation. Thus, rejecting Hypothesis 4a, 4b, and 4c.

4.4 Robustness analysis

(28)

28 and the moderating effect of HTS on that relationship. The subgroup US + UK + AUS shows similar results in comparison to the results of the regression using the entire sample. The results of the other subgroups and the US show no significant evidence of a relationship between INSTITOWN and R&D.

Table 9. Ownership concentration and innovation: Geographical subsamples

US

US + UK + AUS

All - US

All - (US + UK + AUS)

Model 13 Model 14 Model 15 Model 16

R&D R&D R&D R&D

OWNCON 0.00005** 0.00003 0.00000 -0.00001 (0.000) (0.000) (0.000) (0.000) ROA -0.06481*** -0.05692*** -0.01817*** -0.01820** (0.003) (0.020) (0.005) (0.007) SALESGRO 0.00127* 0.00047 -0.00083*** -0.00080** (0.001) (0.001) (0.001) (0.000) ASSETS -0.01201*** -0.00764*** -0.00192*** -0.00125** (0.001) (0.002) (0.0001) (0.001) MTB 0.00170*** 0.00212*** 0.00099** 0.00062** (0.000) (0.001) (0.000) (0.000) LEV -0.00424** -0.00542*** -0.00479*** -0.00565*** (0.002) (0.001) (0.002) (0.002) GDP 0.03272*** 0.01132*** 0.00389*** 0.00354*** (0.004) (0.002) (0.001) (0.001) HTS 0.06800*** -0.01619*** -0.01537*** -0.01121*** (0.002) (0.005) (0.004) (0.003) HTS * OWNCON 0.00007** 0.00005** -0.00005 -0.00006 (0.000) (0.000) (0.000) (0.000) Constant -0.14040** -0.00600 0.00034 -0.00603 (0.040) (0.016) (0.006) (0.006)

Country FE No Yes Yes Yes

Industry FE Yes Yes Yes Yes

Year FE Yes Yes Yes Yes

Adjusted R2 0.2143 0.5310 0.4002 0.4448

Observations 15,013 20,246 24,725 14,492

(29)

29

Table 10. Institutional ownership and innovation: Geographical subsamples

US

US + UK + AUS

All - US

All - (US + UK +

AUS)

Model 17 Model 18 Model 19 Model 20

R&D R&D R&D R&D

INSTITOWN 0.00001 0.00002* 0.00000 -0.00001 (0.000) (0.000) (0.000) (0.000) ROA -0.06485*** -0.05692*** -0.01807*** -0.01802** (0.003) (0.020) (0.005) (0.007) SALESGRO 0.00129* 0.00046 -0.00083*** -0.00081** (0.001) (0.001) (0.001) (0.000) ASSETS -0.01200*** -0.00737*** -0.00190*** -0.00123** (0.001) (0.002) (0.0001) (0.001) MTB 0.00171*** 0.00215*** 0.00097** 0.00060** (0.000) (0.001) (0.000) (0.000) LEV -0.00439** -0.00552*** -0.00480*** -0.00569*** (0.002) (0.001) (0.002) (0.002) GDP 0.03326*** 0.01112*** 0.00400*** 0.00370*** (0.004) (0.002) (0.001) (0.001) HTS 0.67426*** -0.01540*** -0.01513*** -0.01073*** (0.002) (0.005) (0.004) (0.003) HTS * INSTITOWN 0.00008 0.00012** -0.00011 -0.00012 (0.000) (0.000) (0.000) (0.000) Constant -0.14786*** -0.00710 0.00141 -0.00857 (0.039) (0.015) (0.006) (0.006)

Country FE No Yes Yes Yes

Industry FE Yes Yes Yes Yes

Year FE Yes Yes Yes Yes

Adjusted R2 0.2128 0.5321 0.3972 0.4448

Observations 15,013 20,246 24,725 14,492

Note: this table reports the estimated coefficients from regression firm-level variables, and firm-level moderators, on R&D in four different subsamples: US, US+UK+AUS, all countries – US, and all countries – (US+UK+AUS). Definitions of variables are provided in the methodology. Model 17 employs year and industry fixed effects, and Models 18, 19, and 20 employ country, year and industry fixed effects; these are not shown in the regression but can be asked for. The regressions use robust standard errors corrected for clustering at the country-level. The sample period is 2002-2013. Standard errors are shown in the brackets. *,**,*** stand for the statistical significance levels of 10%, 5%, and 1% respectively.

(30)

30 significant effect of INSTITOWN on R&D, and the moderating effect HTS on the relationship between INSTITOWN and R&D remains positive and significant. At last, although shareholders’ rights itself hurt R&D, there is no significant influence on the relationship between ONWCON and R&D or INSTITOWN and R&D.

Table 11. Ownership concentration and innovation, institutional ownership and innovation: R&D>0

Model 21 Model 22 R&D R&D OWNCON -0.00003 (0.000) INSTITOWN -0.00006 (0.000) ROA -0.09850*** -0.09830*** (0.009) (0.009) SALESGRO -0.00164* -0.00169* (0.001) (0.001) ASSETS -0.01583*** -0.01575*** (0.003) (0.004) MTB 0.00351*** 0.00353*** (0.001) (0.001) LEV -0.00823*** -0.00831*** (0.002) (0.002) GDP 0.01238*** 0.01249*** (0.002) (0.002) HTS 0.01902* 0.01742 (0.011) (0.012) HTS * OWNCON -0.00002 0.00012*** (0.000) (0.000) SHARE -0.00304** -0.00307*** (0.001) (0.001)

SHARE * OWNCON 1.38e-06

(0.000)

SHARE * INSTITOWN 0.00001

(0.000)

Constant 0.13563*** 0.1323***

(0.049) (0.049)

Country FE Yes Yes

Industry FE Yes Yes

Year FE Yes Yes

Adjusted R2 0.5083 0.5089

Observations 18,632 18,632

(31)

31

Conclusion

Innovation is becoming a key factor for not only organizations, by enhancing their competitive advantage and thereby providing stability and prosperity, but also to countries as it can generate long term value and foster economic growth.

Innovation, however, is not easily achieved for firms. According to Holmström (1989), innovations are risky, unpredictable, long-term, labor-intensive, and idiosyncratic, all creating hurdles for organizations that want to invest in innovative activities. Next to being labor-intensive, innovations require multiple resources such as materials and capital, but also non-materialistic resources as knowledge and a stable political climate (Klein, and Sorra, 1996). The level of institutional ownership can positively influence an organization’s level of innovation due to its influence on management (Holmström, 1982; Choi et

al., 2012) and the resources it introduces to the organization (Chen et al., 2013).

Recent literature in this field predominantly focuses on one country (Choi et al., 2012) or a group of countries such as developing countries (Chen et al., 2013), missing country-specific characteristics and not accounting for different levels of institutional ownership and ownership concentration in different countries. This thesis, therefore, contributes to the literature as it uses a sample of 48 countries in all continents and uses a more diversified theoretical framework by combining the agency theory with the resource dependent theory.

(32)

32 result is robust for a subsample consisting of the US, the UK, and Australia, and also a sample consisting of only observations with R&D values of higher than zero.

Conversely, the sample analyzed does not find a significant relationship between ownership concentration and R&D, and the moderating effect of firms operating in high technological sectors is negative. Although the level of shareholders’ rights in a country directly affect innovation, it has no significant influence on the relationship between institutional ownership and R&D or ownership concentration and R&D.

This thesis is prone to limitations. Due to the unavailability of data, this thesis uses the input of innovation, R&D, as a proxy for innovations. However, the length of particular innovation projects for companies is not known, and therefore, the causality between the independent variables and

innovation could be biased. Also, this thesis follows Francis et al. (2011) in assigning a zero for missing variables for the variable R&D. Thus, further research can extend the validity of the results by gathering more specified data on R&D. Another option would be to use another proxy for R&D, such as patents. Lastly, some noteworthy results came forward in the subsample US in the robustness analysis of this thesis. Further research could explore the factors that contribute to these outcomes.

Acknowledgment

First, I want to express my sincere gratitude to my supervisor, Adri de Ridder, who always

offered me the help I needed to finish this thesis. Next to this, he gave me the confidence

that I had the capabilities to finish this thesis. For me, this balance between help and

(33)

33 References:

Aghion, P., van Reenen, J., Zingales, L., 2013. Innovation and Institutional Ownership. American Economic Review 2013, 103(1): 277–304

Choi, S. B., Lee, S.H., Williams, C., 2011. Ownership and firm innovation in a transition economy: Evidence from China. Research Policy Volume 40, Issue 3, April, Pages 441-452

Francis, J., Smith, A., 1995. Agency costs and innovation Some empirical evidence. Journal of Accounting and Economics 19 (1995) 383-409

S. Grossman, O. Hart, (1980) Takeover bids, the free-rider problem, and the theory of the corporation. Bell Journal of Economy, 11, pp. 42-64

Holmström, B., 1989. Agency costs and innovation. Journal of Economic Behavior & Organization Volume 12, Issue 3, December 1989, Pages 305-327

Klein, K.J. and Sorra, J.S., 1996. The challenge of innovation implementation. Academy of Management Review. Volume 21, No. 4

Deng, Z., Hofman, P.S. and Newman, A., 2013. Ownership concentration and product innovation in Chinese private SMEs. Asia Pacific Journal of Management. September 2013. Volume 30, Issue 3, p.717-734

Baccara, M., Razin, R., 2009. Innovation and corporate conservatism (Unpublished working paper). Washington University in St. Louis and London School of Economics.

Acharya, V.V., Baghai, R., Subramanian, K., 2014. Wrongful discharge laws and innovation. Review of Financial Studies 27, 301–346.

Porter, Michael. 1992. “Capital Disadvantage: America’s Failing Capital Investment System.” Harvard Business Review 70 (5): 65–82.

Kaplan, Steven, and Bernadette Minton. 2006. “How Has CEO Turnover Changed? Increasingly Performance Sensitive Boards and Increasingly Uneasy CEOs.” Unpublished

Hart, Oliver D. 1983. “The Market Mechanism as an Incentive Scheme.” Bell Journal of Economics 14 (2): 366– 82.

Pfeffer, J. 1972. Size and composition of corporate boards of directors: The organization and its environment. Administrative Science Quarterly, 17: 218–228.

Corporate structure, ownership and the likelihood of innovation. James H. Love , Brian Ashcroft & Stewart Dunlop. Pages 737-746. Journal Applied Economics 2010

Journal of International Business Studies. March 1991, Volume 22, Issue 1, pp 41–62.Strategic Determinants of Partner Selection Criteria in International Joint Ventures. J. Michael Geringer.

Journal of World Business. Volume 38, Issue 1, February 2003, Pages 15-30. International joint ventures: creating value through successful knowledge management. Iris Berdrow Henry W Lane

Effects of Foreign Ownership On Innovation Activities: Empirical Evidence for Twelve European Countries. Martin Falk. First Published April 1, 2008 Research Article. National Institute Economic Review

Asia Pacific Journal of Management. September 2013, Volume 30, Issue 3, pp 717–734. Ownership concentration and product innovation in Chinese private SMEs. Ziliang Deng, Peter S. Hofman, Alexander Newman

Hall, L. A. and Bagchi-Sen, S., 2002. A study of R&D, innovation, and business performance in the Canadian biotechnology industry. Technovation (Elsevier) Volume 22, Issue 4, April 2002, Pages 231-244

(34)

34 O’Sullivan, D., & Dooley, L. (2009). Applying Innovation. (I. SAGE Publications, Ed.). United Kingdom

Alfaro, L., Chandab, A., Kalemli-Ozcan, S.,Sayek, S., 2004. FDI and economic growth: the role of local financial markets Laura Journal of International Economics 64 p.89 – 112

Seifert, B., Gonenc, H., 2012. Creditor Rights and R&D expenditures. Corporate Governance: An International Review, 2012, 20(1): 3–20.

Francis, B. B., Hasan, I., Sharma, Z., 2011. Incentives and innovation: Evidence from CEO compensation contracts. Unpublished working paper. Bank of Finland Research Discussion Paper No. 17/2011.

Chang, X., Fu, K., Low, A., Zhang, W., 2015. Non-executive employee stock options and corporate innovation. Journal of financial economics, 115(1), 168-188.

Djankov, S., McLiesh, C., Shleifer, A., 2007. Private credit in 129 countries. Journal of financial Economics, 84(2), 299-329.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1998. Law and finance. Journal of Political Economy 106, 1113–115

Lee, P. M. & O’Neill, H. M. 2003. Ownership structures and R&D investments of U.S. and Japanese firms:Agency and stewardship perspectives. Academy of Management Journal, 46: 212–225.

Cohen, W. M., & Levin, R. C. 1989. Empirical studies of innovation and market structure. Handbook of Industrial Organization, 2: 1059–1107.

KOCHHAR, R. and DAVID, P.,1996. INSTITUTIONAL INVESTORS AND FIRM INNOVATION: A TEST OF COMPETING HYPOTHESES. Strategic Management Journal, Vol. 17, 73-84

Jensen, R., Showalter, D., 2004. Strategic debt and patent races. International Journal of Industrial Organization, 22(7), 887-915.

Kwon, S. S., Yin, Q. J., 2006. Executive compensation, investment opportunities, and earnings management: High-tech firms versus low-tech firms. Journal of Accounting, Auditing & Finance, 21(2), 119-148.

The impact of corporate governance practices on R&D efforts: a look at shareholders’ rights, cross-listing, and control pyramid. Stephane Lhuillery. Industrial and Corporate Change, Volume 20, Issue 5, October 2011, Pages 1475–1513,

Academy of Management JournalVol. 32, No. 2. Diversification Strategy and R&D Intensity in Multiproduct Firms. Barry Baysinger and Robert E. Hoskisson. 2017.

Volume 65, Issue 3, September 2002, Pages 337-363. Journal of Financial Economics. Funding growth in bank-based and market-based financial systems: evidence from firm-level data. Aslı Demirgüç-Kunt Vojislav Maksimovic.

Successful industrial innovation: critical factors for the 1990s. Roy Rothwell. Volume22, Issue3. July 1992 Pages 221-240. R&D Management.

Institutional Factors in Information Technology Innovation. John Leslie King, Vijay Gurbaxani, Kenneth L. Kraemer, F. Warren McFarlan, K. S. Raman, C. S. Yap. Published Online:1 Jun 1994. Volume 5, Issue 2 June 1994 Pages 75-190. Information Systems Research.

Research Policy. Volume 28, Issue 8, November 1999, Pages 891-905. A resource-based analysis of the factors determining a firm's R&D activities. Jesús Del Canto, Isabel González

FINANCIAL INNOVATION AND ECONOMIC PERFORMANCE Robert C. Merton. First published: Winter 1992. Volume4, Issue4. Winter 1992. Pages 12-22. Journal of Applied Corporate Finance.

(35)

35 Journal of Development Economics. Volume 55, Issue 1, February 1998, Pages 133-153. International intellectual property rights protection and the rate of product innovation. Lai, E.L.C.