Corporate governance, insider ownership and risky investments in R&D projects: developed and emerging markets

Размещено на http://www.allbest.ru/

NATIONAL RESEARCH UNIVERSITY HIGHER SCHOOL OF ECONOMICS

DEPARTMENT OF ECONOMICS

BACHELOR'S THESIS

Corporate governance, insider ownership and risky investments in R&D projects: developed and emerging markets

Student: Anna Tereshchenko

Scientific Supervisor: Anastasia Stepanova

Contents

Introduction

1. Literature review

2. Data Detail, Methodology and Empirical Results

3. Potential Explanation of Our Main Results

4. Discussion

Conclusion

References

Appendix

Introduction

Investment policy is very important for high-tech companies. For such companies R&D investments are one of the major drivers of successful development and enterprise value creating. The most important specific feature of R&D investments is the fact that their outcome cannot be predicted in a precise way. Because of the high risk R&D investments depend not only on fundamental factors, but also on management incentives and attitude toward risk. Management decisions can be influenced not only by their personal characteristics, but also by corporate governance mechanisms and shareholders' intervention. That is why the investigation of relationship between corporate governance, insider ownership and risky R&D investments is extremely important for understanding of corporate investment policy.

Since Jensen and Meckling (1976) research, agency problem is often discussed in terms of situations where managers' incentives do not correspond to shareholders' interests. Studies conducted by Brenner S. (2015) shows that managers often try to avoid risk, Brochet et al (2015) show that managers planning horizon is usually short-term. Lu and Wang (2015) show that if managers are not sure about the outcome of the project, they would rather avoid it to save their reputation and position. Such evidence advocate that corporate governance mechanisms and shareholders' monitoring is needed to prevent investments from distortion.

In this paper we examine how corporate governance and insider ownership influence corporate R&D investments in developed and emerging markets. We analyze data from pharmaceutical and biotechnological industries because R&D investments are extremely important in this sector. We expect the effect of corporate governance and insider ownership to be more significant in emerging markets because in such markets the process of weaker institutional regulation and absence of well-established investment decision-making process.

The object of the study is determinants of R&D investments in risky industries; the subject of the study is influence of corporate governance and insider ownership on risky R&D investments.

Such research is extremely relevant and has a lot of possible policy implications. Firstly, it gives the evidence on the effect of corporate governance effect on R&D investments. That can be the foundation for recommendations about Board composition. Secondly, it shows whish proportion of insider ownership is more advantageous in terms of corporate R&D investments. Thirdly, it can be the basis of implementation of government regulation of corporate governance characteristics, such as minimum (maximum) percentage of independent directors.

The structure of the paper is following. In Section I we present existing literature review, in Section II we reveal data detail, methodology and empirical results. In Section III we propose potential explanation of our main results. Section IV presents conclusions.

corporate financial investment insider

1. Literature review

Risky R & D investments influence firm's performance. This factor is vital for highly technological firms, as their success depends on their technological advantage. Due to significance of R&D investments scholars try to determine factors that can influence investments and innovation activity of companies.

Existing literature considers two major groups of factors that influence corporate investment process: financial determinants and behavioral factors. Financial factors are important as they describe objective characteristics that influence corporate investment activities. Behavioral factors are also important because that reflect incentives of the parties that can influence investment process.

Financial factors

Financial factors are usually analyzed in terms of the corporate structure that is more favorable for large projects financing. The most common factors that can influence investment activity are factors that define budget constraint.

Firstly, that is free cash flow that the firm generates. A large number of studies, starting from Fazzari, Hubbard, and Petersen (1988) suggest that there is positive effect of higher free cash flow on investment activity. In case of perfect markets there should be no tight connection between the internally generated cash flow and investment volume, but in reality companies usually have significant costs of raising the money and placement of free funds on the external capital market. Such market inefficiency was described by Richardson S. (2006). According to the Richardson research, firms with poor free cash flow tend to underinvest, while companies with significant free cash flow tend to overinvest. Gupta and Bhatia (2016) investigated Indian companies for the period 2004-2012 and showed that both free cash flows and cash holdings positively influence corporate investments. Alti (2003) proposed that free cash flow not only serve as budget constraint proxy, but also signals of investment opportunities, as successful companies with higher free cash flow have more professionalism and connections that broaden possible investment set.

Secondly, cost of financing matters while making investment decision. If the company can attract funding with minor costs, it can accept the projects with moderate rate of return. Frank and Shen (2016) empirically demonstrated that weighted average cost of capital is negatively related to corporate investment. The regression model showed that higher implied costs of capital are associated with fewer investments. This finding is in line with research conducted by Richardson S. (2006). However, for companies from volatile sectors, such as pharmaceuticals and biotechnology, WACC is not only shows the cost of capital, but also can be used as proxy for company's current and forward risk. For highly volatile companies high WACC highlights risky strategy that implies higher risky investments. Such effect of forward-looking WACC interpretation was proposed by Lorenz, D., Kruschwitz, L., Loffler, A. (2015). In their research they demonstrated that future plans and strategy should be taken into consideration and WACC can be treated as proxy for “risk appetite”.

Another important factor is dividend payout ratio. Existent literature proposes that high payout ratios create additional budget constraint and negatively influence investment process. Crisуstomo, V.L., Iturriaga, F., Gonzбlez, E (2014) presented empirical evidence of negative connection between dividend payout and investment level in the emerging markets using the sample of non-financial Brazilian companies. Similar empirical results for Korean companies were achieved by Lee M. and Hwang (2003). Their research showed that in the regression analysis of R&D investments the coefficient for dividends is negative and significant. This finding shows that high dividends can restrict cash flow and negatively influence investment level.

Lee M. and Choi M. (2015) analyzed financial determinants of the corporate R&D investments in pharmaceutical industry. They investigated the influence of liquidity ratio, debt ratio and ROI on the R&D intensity of Korean pharmaceutical companies during the period from 2000 to 2012. Linear regression analysis showed that the liquidity ration positively influence R&D investments, while debt ratio has negative effect on R&D intensity. Such result is in line with the expectations that companies with higher liquidity have more resources to finance R&D process, while the companies with high leverage have fewer opportunities to attract additional capital to finance more projects. The authors have not revealed significant connection between ROI and R&D investments. The explanation of such result is based on specific features of the industry, where clinical-stage companies do not generate any revenue and imply significant loss. Thus, investment process is forward-looking and current ROI does not determine R&D budget.

Behavioral factors

Behavioral factors are extremely important considering R&D investments. Such investments are very risky and the outcome is extremely hard to predict. The main parties involved in the investment-decision process are managers and shareholders.

Managers Incentives and Board Composition

It is important to note that managers have their own incentives which can be different from shareholders' interests. Research conducted by Brenner S. (2015) proposes the empirical evidence that managers often try to avoid risk. The option exercising data (1996 to 2008) was taken as proxy for risk aversion in the regression analysis. Another important concept is planning horizon. Considering the managers planning horizon the concept of «managerial myopia» is usually used. This term describes that short-term preferences are more important for managers. Brochet. F, Loumioti. M, Serafeim. G (2015) studied the conference calls with executive managers and disclosure channels and determined that the myopic behavior is typical for managers. Chowdhury J, Sonaer G (2015) constructed theoretical model of management myopia which shows that managers prefer short-term profit to firm-value maximizing behavior. One of the reasons that intensify such effect is compensation on the current financial results. Such system creates incentives to decrease the risky R&D expenses that are not necessary for short-term operations. That improves current financial results and minimizes the risk of negative outcome in the future that can negatively influence managers' reputation. According to such managerial incentives and information asymmetry agency problem arises, which was described by significant segment of scientific research (Holderness et al., 1999; Kole, 1995; Morck et al., 1988; Schmidt, 1975 etc.). Existing literature showed that managers are not diversified and more risk-averse than shareholders. (Amihud and Lev, 1981; Hirshleifer and Thakor, 1992). Managers try to protect their reputation and wealth and avoid risky investment that can have negative outcome. Ben-Zion (1984), Bhagat and Welch (1995) and Kothari et al (2002) showed that managers myopia and risk-averseness increased with uncertainty of the project's outcome.

Moreover, uncertain R&D projects which are hard to evaluate imply additional efforts to analyze. Thakor (1993) research shows that managers try to avoid such investments that are associated with more efforts, high risk and less current benefits.

Prior research has also proven that managers use R&D budgets to adjust company's financial result to the target. They try to avoid earnings disappointment, such as earnings decreases (Bushee, 1998) and negative surprises (Bange and De Bondt, 1998). Graham, Harvey, and Rajgopal (2005) survey shows that over 80% managers would rather cut R&D investments to get the target financial results, 55.3% managers would simply delay the R&D project run.

High probability of occurrence of agency problems requires the establishment of mechanisms for its mitigating. Such mechanisms can include independent directors in the Board and CEO duality

A. Proportion of independent directors in the Board

Independent directors have no material or personal incentives in the company (Luoma and Goodstein, 1999). Their incentives are supposed to be close to shareholders' incentives. Thus, the higher proportion of independent directors in the Board should mitigate described above managers-shareholders agency problem. Existing literature shows that both positive and negative effects of independent directors can exist.

Positive effect is mostly associated with the fact that independent directors can objectively estimate the projects and do not require additional risk premium. They do not warry about financial targets, bonuses and their position. Dong and Gou (2010) showed that increasing number of independent directors leads to more R&D investments. Osma (2008) also studied independent directors - R&D investment relationship and proposed empirical evidence that independent directors can mitigate opportunistic behavior of managers and reduce opportunistic decreases of R&D investments.

Another studies show that independent directors can negatively influence R&D investments. The main reason of such effect is the lack of information accumulated by independent directors. Gilson & Kraakman (1991), Hill & Snell (1988), Hoskisson, Hitt, Johnson and Grossman (2002) propose that independent directors have extremely high costs of information gathering and rely only on financial data on investment projects. Due to the fact that they have no company specific understanding of investment process they distort R&D investments. Beekes, Pope, and Young (2004) formulated conditionals that are necessary for positive influence of independent directors on the corporate investment process. The first one is incentives to monitor the managers' decisions. The second condition is sufficient knowledge that allows to understand investment opportunities and professionally correct management plans. The third condition is recognition that managers have incentives to underinvest it R&D projects to avoid the risk and improve financial result. Such understanding motivated to formulate personal objective opinion, which is not based on management point of view. Bhagat and Black (1999) and Peasnell et al. (2000) highlight that the second condition is often hard to meet, and without it independent directors do not improve investment process. Adams and Ferreira (2007) propose that additional monitoring from independent directors is hostile to managers and they will less likely rely on the Board advice. The critical case is not friendly management board limiting information for independent directors and increasing information asymmetry that negatively influence investment process.

Recent studies on corporate governance in high-tech shows that the first effect is most common. Following the logic of Peasnell et al. (2000), that means that independent directors have sufficient incentives and competence to monitor managers. The result is positive connection between proportion of independent directors and corporate R&D investments.

Lu and Wang (2015) promote empirical evidence that higher proportion of independent directors positively influence R&D investments in the developed markets. They build the regression for USA firms and found out that higher proportion of independent directors increases R&D investments and decreases capital investments. Such results can be treated as evidence that corporate governance mechanisms can mitigate agency problem, which is characterized by overinvestments in tangible assets and underinvestment in R&D. Excess capital expenditures are often explained by managers' incentives to «build the empire» and have more assets under control (Harford, 1999; Lang, Stulz and Walkling, 1991). Decreasing of capital investments and increase of R&D investments in companies with higher proportion of independent directors can be a signal of more optimal investment structure.

Existing literature possesses sufficient empirical evidence that proportion of independent directors positively influence company performance (Liu et al, 2015, Bradley and Chen, 2015, Aras 2015). These results do not directly show that independent directors improve investment efficiency, but these results can be treated as proxy of overall efficiency of such corporate governance feature.

B. CEO duality

CEO duality concept means that one person occupies CEO and Chairman of the Board (Rechner and Dalton 1991). Existing literature on the CEO duality effect can be divided into three groups.

The first group advocates that the separation of CEO and Chairman of the board positively influence investment process. Fama and Jensen 1983 and Jensen 1993 claim that in case CEO also holds Chairman' position the Board lose its monitoring function and agency costs increase. The separation allows the Board to monitor CEO and brings more efficiency to corporate actions and performance. CEO duality contradicts the concept of «сheck and balance» and increases the risk of opportunistic behavior.

In case one person holds both positions, he has more power and more opportunities to promote his decision in the Board. Thus if CEO is opportunistic he will distort corporate investments and maximize his own utility (Fama & Jensen, 1983; Jensen, 1993).

Moreover, if CEO is lazy and indifferent to corporate investments, he will avoid all the costs of searching for new investment opportunities and analyzing projects suggested by other managers. Such behavior was analyzed by Bertrand and Mullainathan 2003 and Aggarwal and Samwick (2006). According to the study such behavior suppresses initiatives in the Board and negatively influence company investment process. Limiting of the CEO power by Chairman monitoring will improve the situation and reduce CEO opportunistic behavior.

Research conducted by Daily and Dalton (1992), Pi and Timme (1993), Baliga, Moyer and Rao (1996) contain empirical evidence of positive effect of CEO - Chairman of the Board separation on corporate performance. Improvement of corporate performance can be treated as efficiency of the concept and we can expect positive effect on investment process.

The second group of research claims that CEO duality is positive factor and CEO - Chairman of the Board separation negatively influences investment efficiency and corporate performance.

The arguments for CEO duality are mostly based on the argument that CEO has specific knowledge on the company's operations and promotes the right decisions. According to Brickley et al. (1997) CEO power in the Board contributes to fast reaction in tough business environment. Dalton et al. (1998) claim that consolidated power possesses determined business strategy that is beneficial for a company. Stoeberl and Sherony (1985) and Anderson and Anthony (1986) also supported CEO duality concept. They argue that CEO-Chairman separation leads to cost of information-sharing and conflicts between CEO and Chairman of the Board. Empirical evidence was proposed by Brickley, Coles and Jarrell (1996). Their research shows that firms with CEO duality perform better.

The third branch of research claims that there is no significant difference between companies with CEO duality and companies with CEO - Chairman separation. Daily and Dalton (1997) empirical research shows that companies with different leadership structures considering CEO duality have no significant differences. Dahya and McConnell (2005) also report that there is no significant advantage for the firms with CEO Chairman separation.

Such controversial results show that there is no clear pattern that connects CEO duality and corporate performance and investment efficiency. Brickley, Coles, & Jarrell (1997) advocate that there is no one optimal leadership structure that brings prosperity to companies. Scientists claim that both CEO duality and CEO - Chairman separation have related advantages and costs. Efficiency of specific leadership structure also depends on the industry and personal qualities of CEO and other managers. Thus CEO duality can be inefficient for some companies and beneficial for others.

Insider Ownership

As in current research, insider ownership is analyzed as proportion of shares outstanding held by managers and beneficial owners.

Positive relationship between managerial ownership and corporate performance arises from agency problem (Lambertini and Mantovani, 2010, Jensen and Meckling, 1976). Managers that do not own stocks have incentives to maximize their own utility regardless the interests of shareholders. Thus, managerial ownership can converge interests of managers and shareholders and positively influence corporate performance. Chung and Pruitt (1996) investigated US companies and found out that executive ownership has positive connection with firm performance. The same results were achieved in investigation of European companies (Kaserer and Moldenhauer, 2008). Investigation conducted on UK companies sample by Crossan (Crossan, 2011) also showed positive connection between insider ownership and performance.

Research conducted by Morck et al. (1988) shows that relationship between managerial ownership and corporate performance is non-linear. The relationship is positive if management own proportion in a ranges of 0-5% of shares outstanding, negative if managers own 5-25% shares outstanding and positive for board ownership above 5%. The scholars describe such effect by entrenched effects that can be observed when managers hold 5-25%. Such effect is characterized by management entrenchment. Managers own a stake of the company that is not big enough to make their interest close to shareholders' ones, but they gain more power that can negatively influence firm's performance.

Analysis of Romanian companies conducted by Vintila and Gherghina (2013) found out negative connection between insider ownership and investment efficiency. Although the authors tested non-monotonic model, they have not registered any positive effect of insider ownership. Vintila and Gherghina explained their results by prevalence of entrenchment effect during observed period (2007-2011). Although the majority of previous research identified positive influence of insider ownership in developed markets, analysis of companies in emerging markets shows that entrenchment effect can prevail.

There is empirical evidence that study non-linear relationship between insider ownership and corporate investments. Davies et al. (2005) investigates US companies and showed that investment process in negatively influenced in case insiders own more than 26% but less than 51% stake of the company. Vary high insider ownership (more than 76%) have also negative effect of investments. Analysis of companies in India shows U-shaped relationship between insider ownership and investment efficiency: investment efficiency is negatively influences if insider ownership is less than 45% - 63%. Subsequent increase of insider' stake positively influence investment efficiency (Selarka, 2005). Similar results for Indian companies were achieved by Pant and Pattanayak (2007). Insider ownership in the interval 0-20% positively influence investment efficiency, 20-49% insider ownership stake have negative influence on investment efficiency, proportion of insider more than 49% have positive connection with investment efficiency.

Nunn et al. (1983) proposed that aggregated number of insider ownership proportion is not necessarily has an impact on corporate performance. Different types of insiders have different abilities to influence investment decisions. Top executives are supposed to have greater power in promoting their view and better access to non-public information. Increasing of stake of such ownership is supposed to have greater effect on corporate performance.

Thus, there is no one clear influence pattern and the effect of insider ownership in high-tech industries should be further investigated.

2. Data Detail, Methodology and Empirical Results

We start our research of the firm's investments from investigation of a decision-making process. R&D investments in pharmaceutical and biotechnology industries are extremely risky due to the fact that the outcome cannot be predicted in a precise way. Different parties have different incentives that are led by risk- averse nature, scientific interest and ambition to invent a product that will generate revenues. The main participants of investment decision-making process are managers, shareholders and scientists.

A. Managers

Managers that make the decisions about investments have motives to invest less in risky projects.

First of all, they want to hold their position. Managers cannot certainly predict the outcome of such projects. Unlike shareholders, which can hold shares of different companies and diversify their risk, managers' wealth depends on the performance of particular company. Unsuccessful investments can be treated as the result of lack of competence, and managers try to avoid such situations (Kothari et al, 2002)

Secondly, a manager's salary often depends on the firm performance. The bonuses can be connected with the revenue for the period. As R&D expanses decrease the revenue and are not necessary for getting result in short-term period, managers don't want to keep them at a high level. Moreover, managers can adjust corporate financial to the plan by cutting R&D expenses (Graham et al., 2005).

Thirdly, the results of R&D projects make a firm prosperous in long-time period. Often some years needed to get the result of inventing a new product or technology. Managers are not sure about their position in long-term period and do not always think about long-run perspectives (Brochet, Loumioti, and Serafeim 2015).

Due to these reasons managers tend to invest less in R&D. Advanced corporate governance is proposed to mitigate the described problem. Percent of independent directors on the Board and CEO duality are considered as main factors that can make decisions taken by managers less distorted.

Proportion of independent directors in the board

Independent directors are members of the board who have no material or personal interest in the company. They do not own the company's stocks and they can impartially evaluate the issues concerning the company (Luoma and Goodstein, 1999).

Considering R&D investments independent directors can fairly estimate the risks of the project and the benefits it can yield (Osma, 2008). As they do not warry about their position, they do not require additional premium for risk to approve a project. Independent directors can insist on undertaken the risky projects in case they seem to be profitable.

The significance of independent directors increases when the risk of the projects increases, because even in high-tech companies managers try to avoid taken enormous risks (Brenner S., 2015). If perspective projects have smaller probability of negative outcome and smaller volatility, they can be undertaken by managers even without independent directors' monitoring. In case risk increases, managers try to hide the investment opportunities which outcome can harm their reputation, but independent directors can understand the necessity for such innovations and promote such projects.

In order to estimate the projects in a correct way, independent directors must have competence in the company's operations and growth perspectives. Getting the competence involve the costs of information gathering and analyzing. Lack of insiders with deep knowledge of the company and the industry in the board increase the costs of information gathering for outsiders. Excess amount of independent directors can harm strategy orientation (Johnson and Grossman, 2002). Thus, excess number of independent directors negatively influences R&D investment. However, the number of independent directors usually not excess, that is why negative effect does not seem to be significant.

Hypothesis 1:

There is positive relationship between proportion of independent directors and R&D investment intensity.

Although we expect positive relationship it is important to include the polynomial variable of proportion of independent directors in the model in order to check whether inverse U-shaped relationship between proportion of independent directors and R&D investment intensity exists.

CEO duality

The concept of CEO duality indicates whether the separation of Chief Executive Officer and the Chairman of the Board exists (Rechner and Dalton 1991).

If one person holds both positions he has more power in promoting decisions and more responsibilities. Existing literature contains mixed results of CEO duality relation with firm performance.

We expect to get negative relationship of CEO duality and R&D investment intensity. The logic of such relation is connected with power and responsibility concentration. If one person supervises all the executive team he is the main risk-taker, and a risk-averse manager can try to avoid risky investment and even hide the information about R&D investment opportunities (Moyer and Rao, 1996).

If duality exists, two people can provide more diverse view about a risky project. In such case CEO is not the one powerful and responsible figure that that can promote and take serious decisions. Presence of additional powerful person in the board implies more discuss on important project. That can reduce the level of information asymmetry and positively influence investment decision-making process.

Hypothesis 2:

CEO duality is negatively associated with R&D investment intensity.

B. Shareholders

Shareholders are interested in long-run company perspectives and fundamental firm value. It is important to understand the profile of peoples that invest in volatile risky companies. They prefer to get high return on the projects which imply significant risk. That means that they would like to have wide research activity that can result in developing the product that can generate significant revenues.

Another important concept is number of shares hold by CEO and other executives. If managers of the company hold the shares, they have their personal interest in both short-term and long-term successful company's performance (Lambertini and Mantovani, 2010). That is why the agency conflict between shareholders and managers decreases. In such case, managers are interested not only in their position and bonuses for the period, but also in company's prosperity (Jensen and Meckling, 1976). Increasing number of shares held by managers implies convergence between the managers' and the shareholders' interests.

Managerial ownership increases the planning horizon, what positively influence risky R&D investments. The planning horizon increases as managers want to receive more benefits of the company long-term success. Managers become personally involved in company's development process (Lambertini and Mantovani, 2010).

Due to the fact that the information on the managerial ownership is not available for the majority of the pharmaceutical and biotechnological companies on emerging markets, it is problematic to get any empirical results. In this research we will analyze the influence of the percentage of insider shares on the R&D investments. Percentage of insider shares includes percentage shares hold by managers and percentage of shares hold by beneficial owners. Beneficial owners are the shareholders that have more than 10% of shares outstanding and have power to influence the decisions in the company.

Percentage of insider shares can be used as a proxy for the proportion of ownership of interested people in the company. This term id different from ownership concentration, but it shows the degree of equity fragmentation and can be used as indicator of percent of people who is highly personally interested in company's performance.

Hypothesis 3:

Percentage of insider shares is positively associated with R&D investment intensity.

C. Scientists

Scientists are always involved in the R&D process in pharmaceutical and biotechnology industries. Majority of the companies have the group of researchers in the company. Some companies additionally cooperate with scientific institutes and research universities. Scientists are people who are interested in technology development and have less interest in revenue generation. That gives additional incentive to invest more funds in R&D. If scientists have a lot of influence on the process of money allocation, they can promote to spend more money on innovation development.

Hypothesis 4:

Scientific connection is positively associated with R&D investment intensity.

It is important to note that the connection with science community can be linked to corporate governance and insider ownership. Institutes and research universities have special knowledge and scientific expertise that can be beneficial for an innovative pharmaceutical and biotechnological company. Thus, a lot of insiders that own share of the company can promote more scientific connections to bring accelerated development. Corporate governance mechanism such as independent directors and CEO duality can bring additional discourse and expertise in the Board. These can be the reasons of better strategy of company's development and more connections with scientists.

Control factors

Free cash flow to the firm (FCF)

This factor reflects the financial opportunity of the company to make investments. Companies with severe budget constraints tend to invest less in their new products due to the fact that they have no resources. Companies that generate sufficient revenues are in better financial position and can allocate their money in R&D projects (Richardson S., 2006). Positive connection is expected.

WACC

This factor can be considered as proxy for the company's risk appetite. WACC shows required return on the firm's capital, which depends on the strategy and risk involved. (Lorenz et al, 2015)

Some researchers consider WACC and cost of debt as budget constrain determinants also (Richardson S., 2006). They propose that companies with cheap financial recourses can easily obtain additional funds to invest them in perspective projects. However, pharmaceutical and biotechnological industries are the cases of extremely different companies with different williness to bear risk. Some companies have a moderate risk strategy and investors perceive them as less risky and require less return on the capital invested. Some companies undertake a lot of risky projects with hardly predictable outcome. Thus the weighted cost of capital reflects the propensity to take risk in order to develop new products.

Positive connection is expected.

EBITDA variance

EBITDA variance is considered as the volatility factor. It reflects the magnitude of financial results' variance. We consider EBITDA variance as backward-looking firm's attitude toward risk. Positive connection is expected.

Sample and Data

To investigate determinants of risky investments we should analyze the companies from the industries that imply a lot of risky R&D projects. Moreover, the effect of corporate governance and insider ownership on the company's investments is expected to be more perceptible in industries where the outcome of investments is unpredictable. For this reasons the healthcare sector was chosen for consideration. This sector contains mainly biotechnological and pharmaceutical companies. Not all the drug-makers were included in the sample. The selection criteria was a presence of own drug development process in a company. Generic companies that produce and commercialize already approved drugs were excluded from the sample, as they do not conduct R&D activities.

The samples contain only public companies. We investigate companies on this life cycle stage because this stage usually involves the main part of product development. Before the approval and commercialization of products, pharmaceutical or biotechnological companies mainly do not generate any revenue, but they need significant resources to spend for R&D. That is why at the stage of active research companies make IPO's and attract shareholders financing to conduct the research. Private companies usually have only early-stage R&D which does not require significant investments.

In order to estimate the effect of corporate governance and insider ownership on the R&D risky investments, we constructed two samples. The first one contains the data for developed markets (USA, Japan) and emerging countries (China, India, Indonesia, Malaysia). We analyze these countries because public pharmaceutical and biotechnological companies with active R&D activities are based there. In other countries high-tech pharmaceutical and biotechnological industries are not developed. The total number of public pharmaceutical and biotechnological companies in developed countries' sample is equal to 517 companies, in emerging countries it is equal to 482 companies.

We studied the operation activity of companies and excluded companies that do not conduct active R&D and just commercialize generic drugs. The samples shortened to 339 companies in developed countries sample, to 294 in emerging sample.

After we collected the available data on corporate governance and excluded companies that do not have available corporate governance data. Final developed countries' sample consists of 259 companies; final emerging countries' sample consists of 215 companies.

The analyzed period includes 2013-2015 years. We use transactional data in constructing regression with lags. That is why some data is taken for 2013 year and some data is taken for 2015 year. Collection of panel data is not possible due to missing information on corporate governance for 2012 year.

Developed countries sample

Corporate governance data was mainly collected from the Bloomberg and Thomson Reuters databases. However, some companies have no information in the databases, and missing information was collected from annual companies' reports. Insider ownership data was collected from Bloomberg database, SEC filings and Thomson Insider Data. In defining insider ownership we used information for the total ownership of managerial stocks and beneficial owners stocks. The reason for that is lack of information of just managerial ownership. Concerning the university connection, the data was hand-collected. We checked whether a company declares the collaboration with Scientific Institutes or Research Universities in its annual report. If company has such declarations, we reported for existence of scientific connection. R&D investment data and control variables were uploaded from Bloomberg database.

Emerging countries sample

The scarce data availability for the emerging markets implies a lot of hand-collected data. The tickers and main fundamental data (R&D investments and control variables) was uploaded from Bloomberg database, after that we used information from companies' annual repost to populate our sample with information on corporate governance, insider ownership and scientific connections.

Summary statistics is presented in the Table I.

We provide summary statistics separately for developed and emerging markets.

Firstly, we should note that in the developed markets more pharmaceutical and biotechnological companies have connection with scientific communities. Mean value of dummy variable for scientific connection is equal to 0.667 for the US and Japan, while for the emerging countries mean value for scientific connection is equal to 0.481. High standard errors for both developed and emerging countries shows that the values are not grouped around average value and have high variance.

Secondly, the markets differ by their corporate governance features. Companies in the developed markets are characterized by higher proportion of independent directors. The mean value for percent of independent directors in the board is equal to 76.19% for developed markets, while in the emerging markets mean proportion of independent directors in the board is only 34.43%. However, fewer companies in the developed markets have CEO duality. The mean value of the dummy variable for CEO duality is equal to 0.31 in the developed markets, while for emerging markets it is equal to 0.5. High standard errors for percentage of independent directors (16.16 for developed countries and 15.6 for emerging countries) and high standard errors for CEO duality (0.46 for developed countries and 0.5 for emerging countries) show that the he variance of the corporate governance characteristics is rather high.

Thirdly, the mean value of assets differs more than ten times. For the first glance the fact that companies in the emerging countries have more assets (mean $47779 for developed countries, $118958 million for emerging countries) seems to be strange. But the deeper analysis of balance sheets of biotechnological and pharmaceutical companies shows that inflated assets are typical for such firms in emerging markets. The reason is high cash inflows from collaborations to finance operations in the future. Such companies at clinical development stage do not generate any revenues, but have high R&D expenses. That is why they try to find the partner to finance the research. Actually, such companies usually sell the rights to commercialize the drug but retain rights to get royalty payments. That is why they archive high cash amount in advance (as well as contingent liabilities and deferred revenue in the passive side of balance sheet). Companies in the developed markets do not usually carry such extra cash and they typically raise money by follow-on or debt according to the need.

Another significant difference is the FCF for 2013. In the developed markets it is positive and equal to $2614 million. In the emerging markets FCF 2013 is negative and equal to -$43.36 million. Such phenomenon can be explained by the fact that in the USA and Japan there are a lot of pharmaceutical and biotechnological companies that have approved drugs and generate high revenues. In the emerging markets, in the contrary, overwhelming majority of the companies in the sector are in clinical stage and generate losses at current development stage.

Concerning WACC, we can conclude that on average the capital is cheaper for companies in developed markets (9.6% for developed countries, 10.44% for emerging countries). This fact is in line with higher risk and higher required rate of return for developed markets.

The proportion of insider ownership is higher for the emerging markets. The mean value is equal to 4.87% for developed markets and to 14.8 in the emerging markets. On average, companies in the developed markets spend on R&D projects more, than in the emerging markets (mean R&D is equal to $4497 million in developed countries, $2471 million in emerging countries). The R&D intensity is also higher in the developed than in emerging countries (0.303 in developed, 0.021 in emerging). It is important to note that the standard error is very high for R&D expenses and R&D intensity. That indicated the high variance of the values.

Capital expenditures are significantly higher in the emerging countries ($6762 comparing to developed countries ($759.42). That finding is in line with the fact that the industry appeared in the developed countries first, and a lot of capital investments have been already made. In the emerging countries the industry is nascent, and 2013 was the year of intensive investments.

Table I. Descriptive statistics

	Developed countries		Emerging countries
Variable	Mean	Std. Dev.	Min	Max	Mean	Std. Dev.	Min	Max
SCIEN_CON	0.667	0.472	0	1	0.481	0.501	0	1
PERC_IND_DIR	76.19	16.16	0	100	34.4	15.6	0	100
CEO_DUALITY	0.31	0.46	0	1	0.502	0.501	0	1
ASSETS	47779	328334	0.75	4296192	118958	950002	68.8	1.24E+7
EBITDA_VAR	7.77E+19	1.06E+21	1.53E+11	1.70E+22	7.44E+20	1.00E+22	0	1.47E+23
FCF2013	2614	18494	-7473	249128	-43.356	16938	-172953	168457
WACC	9.6	3.04	0.254	25.053	10.44	2.35	-2.28	18.16
INS_OWN	4.87	8.19	0.01	60.069	14.8	20.8	0	79.53
Q_TOBIN	4.94	10.74	0.701	164.716	3.30	2.31	0.28	13.47
RESEARCH	4497	31572	0.3	386800	2483	28717	0.19	423503
R_D_INT	0.303	0.25	0.006	1.364	0.021	0.018	7.83E-05	0.095
CAPEX	759.42	4801	0	51223	6793	58371	0.02	750705
CAPEX_INT_EC	0.018	0.029	0	0.27	0.055	0.041	0.0001	0.24

The table presents mean, standard deviation, minimum and maximum value of main variables. It is constructed separately for developed and emerging countries. Statistics on percentage of independent directors, CEO duality and insider ownership is calculated on 2014 data, FCF on 2013 data. EBITDA variance statistics was calculated for 2011-2016 period. Other variables statistics is for 2015. The Model

To evaluate the influence of corporate governance mechanisms and insider ownership we use regression analysis. The model is as following:

Where is the dependent variable is - expanses for R&D 2015, independent variables: - lagged proportion of independent directors in the board, - lagged dummy variable for existence of CEO - Chairman separation, - dummy variable that is equal 1 if the collaboration with research universities or scientific institutions exists, - variable that shows the proportion of shares outstanding held by managers and beneficial owners, is lagged free cash flow to firm, is weighted average cost of capital, is measure of EBITDA volatility which was calculated for 5 - year horizon.

The table with variables' description is presented in Appendix I (Table II).

We revised the literature on dividends' sufficiency in investment-decision process, but overwhelming majority of the companies in our emerging countries sample do not pay dividends and we extracted this factor from our analysis.

We also tried to include polynomial form of variables (proportion of independent directors, insider ownership) in the model in order to check whether inverse U-shaped relationship exists. Additional variables were not significant and we excluded them from our analysis.

A. Number of lags

To determine the needed lags we investigated the investment programs of the companies from the sample to determine the time gap between decision about investments in risky R&D project and the moment or real research with costs recognition. The average lag for the pharmaceutical and biotechnological industry is close to one year. That is why the values are taken for different time periods.

The decision-making process is following: the management decide which projects to finance according to the previous year financial results (FCF 2013), cost of debt (WACC), board control (proportion of independent directors, CEO duality), owners preferences and opinion of scientists (scientific connection). That is why the lag for FCF is equal to 2, the lag for proportion of independent directors, CEO duality, insider ownership and scientific connection is equal to 1. EBITDA variance is calculated for 2011-2015 years period. We acknowledge that it would be better to calculate EBITDA variance for 2009-2013 year period, but the industry is emerging and have no long data history. As EBITDA 2011-2015 is used just as a proxy for risk and the companies in the sector do not change their risk-attitude, it can be included in the model. WACC data is not available for 2013 that is why we use WACC for 2015. Legitimate reason for such approach is the duration of the process of product development. According to Association of Clinical Research Organizations (ACRO), average process of drug development is more than 5 years, that is why we assume that the level of risk and required rate of return does not change significantly during this time.

B. Model specification.

In order to evaluate the effect of corporate governance characteristics and insider ownership on the R&D investments we use econometric analysis. We use OLS linear model, linear model with robust option, quantile model (median) and robust model with Huber iterations.

OLS linear model minimizes the difference between observations and predicted value that has linear form.

Due to heteroscedasticity problems, we also constructed linear model with robust option. The normality test (Appendix II) shows that errors are not distributed normally, and that is one of the reasons to use median regression.

As R&D intensity has asymmetric distribution (Appendix V, Appendix VI), it is important to construct median linear model. Due to the fact that the sample contains outliers and R&D intensity variance is rather high, median and mean value are considerably different. It is important to have similar effect of independent variables on both for mean and median value of the R&D intensity to draw any conclusion on consistent patterns.

Robust model with Huber iteration is used for samples with outliers. During the estimation the observations are weighted according to its distance to predicted OLS values. Thus observations with high residuals are down-weighted.

The results of estimation are presented in Table III for developed countries and in Table IV for emerging countries. Additional tests for heteroscedasticity, multicollinearity and other possible problems are presented in Appendix III and Appendix IV.

Regressions build on the developed countries sample is significant (p-value = 0.0076). Explanation power of the model is estimated by R² metrics. For linear, robust and median regression R² is close to 8.5 %. This fact means that 8.5 % of R&D risky investment intensity variance can be explained by factors from the model. For the robust model with Huber iterations R² is bigger and equal to 11.7 %. The increase of R² in model with Huber iterations happens because outlier observations are downwardly weighted.

Due to heteroscedasticity concerns, robust regression is a better approach of model estimation. Linear regression and robust regression have minor differences, but they show the same tendency.

...