Luận án tiến sĩ Quản trị kinh doanh: Ownership Structure, Board Characteristics anf Firm Performance in Vietnam

Roles and impacts ofdifferent corporate owners on corporate performance and governance have beenstudied worldwide, but not as popular for the role of State-Owned Holding CompanySOH as a

Research Gap nẽ aắốg “a4 Ỏ

This thesis puts into question much of literature which is adverse to state ownership and suggests to take into consideration the role of SOH The thesis contributes to the limited empirical understanding of how SOH operates as a shareholder in SOH-LinkedCompanies (SLCs) The empirical results differ significantly from adverse argumentation of state ownership in most existing literature Although thesis’s findings could be specific to Vietnam, its knowledge would contribute to a better understanding of the ownership and performance of SOHs around the world Though other studies have been conducted on Vietnamese state capital firms, this thesis not only retests the underlying assumptions of these studies but also includes two completely different aspects First, this thesis analyses in detail the important role ofSCIC, the State-owned Holding Company Secondly, it updates the data used in studies This thesis aims to fill the research gap on SOH by analyzing SLCs that are held by SCIC Although SOH is found in other countries like Singapore and Malaysia,Vietnam is totally different in market classification, financial structure and economic scale to these countries hence the role of SOH may not be entirely homogeneous and suggests to have empirical study to fulfill the gap.

Observations around the world has shown that different owners may exercise dominant roles in different economies, US and UK are dominant by institutional investors, while Asia is dominant by the State and family owners (Nam et al., 1999; La Porta et al.; 1999; Clarke, 2007; Dinga, 2005; Driffield and Pal, 2007) While the role of institutions and family ownerships have been widely studied, the role of State ownership in state-dominant economies have been counted only a few (Bruton et al., 2015) State owners are different from institutional owners or family owners as the State is not driven solely by the benefits of shareholders, but also social benefits, such as unemployment reduction, social group development Also, the state has representatives in corporation who not only pursue the benefit of shareholders but also his or her own private values (Wong, 2004; Wicaksono, 2008; Kamal, 2010; Lin, 2012; Chen, 2013) In some economies where the State exercises a dominant role, there is a model of SOH being established to exercise the ownership and control role in corporations with expectation of professionalism with an exercise of good governance SOH actively manages assets and affect a firm’s management decision with financial return objectives (Al-Hassan et al., 2013) SOH mechanism is expected to mitigate issues of traditional state ownership (Sam, 2013) but studies on SOH is very limited (Sam, 2013; Kim & Chung, 2018) Temasek of Singapore is one among the cases that is called a successful model founded in studies Ang and Ding (2006), Chen (2013) and Kim & Chung (2018) as Temasek bring significant values to the country Though some countries with State dominance also follow this model, few studies discuss the success of the model (Wicaksono, 2008; Chen, 2013; Sam, 2013;

Kim & Chung, 2018) In Vietnam, a country with dominant existence of State ownership (Nguyen et al., 2012) and an introduction of SOH in 2005 with an expectation that this form of ownership will help to mitigate the conflict of interest agency problem and allow the invested-enterprises to achieve good performance but there has not been a specific research on SOH in Vietnamese context to provide empirical evidence yet This study is therefore aiming to examine this type of ownership and its impact to firm performance This study fulfills this need for the context of Vietnam.

1.3 Research Objectives This study aims at achieving the following objectives:

- Investigate the impact of SOH role in managing firms with state capital, measured by the performance of firms with SOH ownership relatively to those without SOH ownerships Given that SOH is a new model of the world and there is extremely limited studies on its role, the findings would be one of the initial contributions to literature of this kind of special state ownership The effectiveness of SOH model in an underdeveloped corporate governance environment would contribute valuable suggestions to policy makers in developing appropriate regulations to support the economy.

- To study the impacts of ownership structure on corporate cash holdings and their interactive effects on firm value Good corporate governance is demonstrated to have positive effect on firm value by improving the value of cash holdings This would be explored in the Vietnamese context but through a new model taking into consideration SOH ownership as a leading role As cash is a neutral asset, the positive impact of SOH role on increasing value of cash holding would yield an evidence of the role of SOH in improving firm value through proper corporate governance.

This study tries to address the following research questions: e Driven by the foremost profitability objective, is the performance of SLCs better than GLCs? e Does SOH help increase firm value through improving value of cash holdings in SLCs?

The scope of this study is limited to listed companies on HOSE and HNX This study focus the discussion on the differences of companies with SOH ownership, companies without SOH ownership, including GLCs and non-GLCs The period of data is from2009 to 2017.

This study uses quantitative research method and applies empirical models that focus on testing the impact of different owners on firm performance, controlling for board characteristics Regression models for panel data are used for testing the explanatory power of variables of interest Various estimation methods are applied to assure the underlying estimation assumptions Specifically, conclusions are derived from the following steps of analysis.

To examine potential differences between the performance of SOH-Linked Companies (SLCs), Government-Linked Companies (GLCs) that has no SCIC ownership and non- GLCs, various market and financial performance measures are examined in univariate analysis, following Ang and Ding (2006) and Chen et al (2006) which highlights the differences between SLCs and others.

The multivariate regressions on firm performance are performed to examine the impacts of ownership structure and board characteristics The model is inherited from previous studies that measure performance by firms’ market values, such as McConnell and Servaes (1990), Short and Keasey (1999), Denis and Sarin (1999), Carter et al (2002), Anderson and Reeb (2003), Ang and Ding (2006), Abdallah &

Ismail (2017), Pillai & Al-Malkawi (2018), Paniagua et al (2018) This study focuses on the ownership of SOH in firms and investigates the positive impact of SOH ownership on firm’s performance.

The power and control of an owner on the firm also depends on its dominant level of its ownership If the owner is relatively the dominant owner, it is expected that the role of the owner is more clearly seen The regression on interaction variables is developed to explore relationship between different types of dominant owners, including SOH ownership, government ownership, family ownership and foreign ownership, and firm performance This model is developed from the study of Chen at al (2006) with elaboration on different types of ownership.

To investigate the impact of better governance in SLCs, models of value of cash holding from previous studies of Opler et al (1999) and Ku et al (2013) are attested.

Also, the value of excess cash with the interaction of ownership level, ownership dominant of different types of owners is studied to investigate the impacts of corporate governance on firm value though the value of cash holding There are two models exploring on level and changes of excess cash which proposed by Faulkender and Wang (2006) and extended by Dittmar and Mahrt-Smith (2007) However, different from these studies, ownership structure, with different owners including SOH ownership, government ownership, family ownership and foreign ownership is added into the model to explore its interaction role with excess cash.

Firstly, Vietnam has characterized by the dominant existence of state ownership in the economy as well as the popularity of state controlling role on stock exchanges This prompts the question for the effectiveness of government’s control on corporate performance As a result, the equitization process is accelerated in Vietnam after many scandals of corruption and poor performance related to State capital firms in recent years Encouragingly, Vietnam is one of the few countries where the model of SOH is applied This thesis contributes to the limited empirical understanding of SOH operates as a shareholder in SLCs and also contribute to a better understanding of the ownership and performance of SOHs around the world This fulfills the research gap on SOH and provide empirical evidences for arguments on the role of SOH as an intermediate mechanism, an institutional investor and an external monitor to improve corporate governance standards.

Secondly, the study also examines how corporate governance impacts a firm’s value through its cash holdings Ownership structure impacts, especially SOH ownership, are examined through cash holdings’ models This is one of first efforts in Vietnam to investigate the relationship between ownership structure, cash holdings and firm value.

This contributes to literature of corporate governance and cash holdings with Vietnamese ownership characteristics.

1.7 Research StructureThe structure consists of five chapters:

Chapter 1: Introduction Introduction of the formation of the subject, objectives, meaning and scope the topic of research.

Chapter 2: State-Owned Holding Company This chapter outlines key concepts of state ownership, state ownership modeling in some countries and, most importantly, the concept of a State-Owned Holding Company with a history of roles and responsibilities of Temasek and SCIC.

Chapter 3: Literature Reviews This chapter would review relevant theories for ownership structure, board characteristics and corporate governance This chapter would also develop hypotheses to be tested.

Chapter 4: Methodology Presentation of methodology, research model and forming measurement variables Regression models are also introduced with analysis methods.

STATE-OWNED HOLDING COMPANY -c <2 23

SOE and SOH nh ae

Given the widespread assumption about the superiority of private ownership over government ownership, a large number of State capital firms have been privatized during the past decades in many economies (Sam, 2008; Wang & Judge, 2012;

Landoni, 2018) State-owned enterprises regularly have poor performance, but there is no well-defined solution to solve problems (Kumar, 1993; Xie & Redding, 2018).

Although the privatization has been accepted as a long-term rational method, not all countries can quickly adopt this solution The transformation of state-owned enterprises requires time and sometimes involves multiple stages (Kumar, 1993).

History shows that duplicating developed countries’ privatization processes would not create same success and unsuccessful attempts can even cause a backlash against the procedure (Wang & Judge, 2012; Birdsall & Nellis, 2005 in Kim & Chung, 2018).

That could be reason to explain why despite extensive privatization across countries, state firms remain persistent in global economies (Kim & Chung, 2018) Policy makers recently turned their attention to the centralized model (World Bank, 2014).

This model allows (1) separating the state's ownership function from policy and regulatory functions to minimize conflicts of interest, (2) minimizing political interference and maximizing professionalism for role (3) improve consistency and consistency to achieve corporate governance standards that play a role in all SOEs and (4) promote transparency and accountability in SOEs operate through effective monitoring and surveillance (Kim & Chung, 2018).

Theoretically, the centralized models can be established through single ministry and state-owned holding company (SOH) forms Under single-ministry system, state firms are organized and overseen by a single specialized entity, such as a ministry of finance An SOH is a parent company owns voting shares of state-invested companies.

This SOH in turn is responsible for a single ministry such as the Ministry of Finance.

SOH does not produce its own goods or services, but its operations include the control and management of the SLCs Thus, SOH acts as the intermediary between government and SLCs Serving as a buffer, SOH in theory enables SLC managers to have greater autonomy than a single set of government systems that can directly affect state capital firms (World Bank, 2014; Kim & Chung, 2018).

State-Owned Holding Company (SOH) is accepted by many developed and developing countries (Kumar, 1993; Kim & Chung, 2018) SOH follows the

“investment company model” which actively manages assets and usually holds the majority of the shares in the company (Al-Hassan et al., 2013; Kim, 2018) SOH manages its own investments to enhance financial returns (Kumar, 1993; Sam, 2013).

SOH allows the appointment of professional managers instead of civil servants into management board (Kumar, 1993; Sam, 2010) SOH also restricts or stops capital transfer from government to its subsidiaries, thus forcing them to improve profitability and efficiency In addition, SOH would stimulate the development of capital markets (Kumar, 1993) In SOH model, the government tries to separate ownership and management and allow the company to operate more freely (Kumar, 1993; Sam, 2008;

Kim & Chung, 2018) This form is continuing widespread (Kim & Chung, 2018) but studies on them are still limited (Sam, 2008; Sam, 2013; Kim & Chung, 2018).

State-Owned Enterprises cccccccccccccccesessssesssnsceeeeeecceeeeessesesssnnaaeeeeeeeeeeeeees 24

According to the Vietnamese business law, only companies with 100% state capital are called state-owned enterprises However, in academic studies, SOEs are a broader concept Therefore, the SOE concept presented here follows the terminology that is widely used in literature.

SOEs are companies operating in commercial lines but owned entirely or partially by a government (Garner, 1970 in Ramamurti, 1987) Willemyns (2016) points out that there is no general definition of the term "SOEs" in international studies and that SOEs are a type of enterprise characterized by distinctive and competitive advantages.

Besides, OECD (2015) states that SOEs are a broad definition and basically, a company is considered as a SOES regardless of whether it is a joint stock company, limited liability company or partnership Entities whose primary purpose is to exercise the rights of the state in general would be considered SOEs Therefore, the characteristics of the SOES presented in the study are cited from previous studies not limited to 100% state-owned enterprises SOEs have been spread rapidly in Western countries and international since War World II as a result of ideological reasons and economic adjustments to promote countries’ developments (Lewin, 1981; Kowalski et al., 2013) The collapse of Soviet Union in 1991 and later mass privatization in Western countries raised concerns on the existence of SOEs (Spicer et al., 2000).

State ownership raises corporate governance matters regarding its specialty in corporate governance and as a result offers variety of theoretical issues for studying (Young et al., 2008) given they are often linked to low efficiency (Bai et al., 2006; Hu et al., 2009; Tan et al., 2015).

Another termimilogy to call SOEs is Government-Linked Companies (GLCs) GLCs are jointly owned by government and private sector These companies are results of partial privatization in many countries It is not clear to determine whether GLC is more efficient or less than private or state-owned enterprises (Sam, 2008) given empricial studies found evidence that GLCs have worse performance than both SOEs

& private firms (Boardman & Vining, 1989; Ehrlich et al., 1994; Oum et al., 2006;

Razak et al., 2011) while Mok and Chau (2003) found that GLCs to be more efficient than private firms although private ones are more profitable.

The state sector still has an important role in many economies, including the most important position (Ramamurti, 1987; OECD, 2005; Kowalski et al., 2013; Chen, 2016; Milhaupt & Pargendler, 2017; Xie & Redding, 2018) However, published studies on SOEs are limited, especially on Financial Times’ Top 45 journals In 15 years from 2000-2014, only 57 journalized articles related to SOEs appeared including two practitioner oriented papers on Harvard Business Review (Bruton et al., 2015).

There is still lacking of studies on SOEs (Bruton et al., 2015) Conflicting objectives, agency issues (political interference) and lack of transparency are considered the main problems of SOEs under pure view of agency theory (Wong, 2004; Kamal, 2010;

Table 2.1 Differences in Governance Between Private and SOEs Sectors

Private sector enterprises State-owned enterprises Objectives Clear focus on capital owners’ value maximization

Pursue commercial and non- commercial objectives

Agency issues Single agency — concerned about self-interested behavior by managers

Double agency — concerned about self-interested behavior by managers and politicians/bureaucrats

Transparency High level of disclosure (for listed firms)

Low level of disclosure Note: Wong (2004), Kamal (2010), Chen (2013), Nurgozhayeva (2017)

Most SOEs pursue multiple — and conflicting — objectives (Wong, 2004; Lin, 2012;

Chen, 2013; Nurgozhayeva, 2017) They are expected to pursue profits but have to guarantee for social responsibilities (Wong, 2004; Chen, 2013; Nurgozhayeva, 2017).

Multiple objectives are arisen because they are mandated by legislation and different ministries might have influence on SOEs simultaneously (Wong, 2004) When a conflict exists between public interests and better corporate governance, the former regularly

Table 2.2 Summary of SOEs related topics on Financial Times Top 45 Journals”

Author(s) Topic Journal Findings Liang et al An anatomy of state control in the globalization of state-owned Journal of International Reduce the impact of administrative political connections as well as increase the (2014) enterprises Business Studies effectiveness of state ownership control over globalization decisions.

Bass and Resource security: Competition for global resources, strategic intent, | Journal of International SOEs collect and pays extra for resources used in exploration rather than Chakrabarty and governments as owners Business Studies exploitation SOEs invest overseas to ensure national security and privacy.

(2014) Choudhury and | Towards resource independence - Why state-owned entities become Journal of International SOEs foster a global footprint and cash flows to achieve resource independence Khanna (2014) multinationals: An empirical study of India’s public R&D Business Studies from home country laboratories Duanmu (2014) | State-owned MNCs and host country expropriation risk: The role of | Journal of International SOEs benefit more than private companies from government protection home state soft power and economic gunboat diplomacy Business Studies regarding host country expropriation.

Li et al (2014) Varieties in state capitalism: Outward FDI strategies of central and local state-owned enterprises from emerging economy countries

Journal of International Business Studies

National champion SOEs bear more institutional pressures from home and host governments than local SOEs.

Overcoming distrust: How state-owned enterprises adapt their foreign entries to institutional pressures abroad

Journal of International Business Studies

SOEs bear more complex institutional pressures in host countries than private companies, acclimating mode and control decisions differently.

Xu et al (2014) Organizational forms and multi-population dynamics: Economic transition in China

SOEs enlarged the exit rate of private companies and collective businesses provided legitimacy for private organisations.

Zeng et al The seller’s perspective on determinants of acquisition likelihood: Journal ofManagement Companies originated as SOEs likely to desire acquisition, unless they have (2013) Insights from China’s beer industry Studies started multiple transformations or involved more private investment.

Inoue et al Leviathan as a minority shareholder: Firm-level implications of state | Academy ofManagement SOEs with small state ownership are less subjected to agency distortions.

(2013) equity purchases Journal Chen et al Are stock option grants to directors of state-controlled Chinese firms | Accounting Review Red-chip SOEs endorsed stock options to directors then forced them to surrender (2013) listed in Hong Kong genuine compensation? options Non-genuine compensation; such options did not affect firm behaviour.

Boubakri et al The role of state and foreign owners in corporate risk-taking: Journal of Financial State ownership is negatively associated to risk taking behaviour in undergoing (2013) Evidence from privatization Economics privatization.

Zeng et al Factors that drive Chinese listed companies in voluntary disclosure Journal of Business Ethics Examines SOEs as a control variable for disclosure of environmental (2012) of environmental information information.

Ke et al (2012) Hong Kong stock listing and the sensitivity of managerial compensation to firm performance in state-controlled Chinese firms

Considers the sensitivity of managerial compensation, effect of long-term incentives and CEO turnover to performance of SOEs.

Du et al (2012) Influence activities and favouritism in subjective performance evaluation: Evidence from Chinese state-owned enterprises

Accounting Review Examination on political connections and geographic nearness to SASAC influence SOES performance evaluation.

Factors impacting ethical behaviour in a Chinese state-owned steel company

Journal of Business Ethics The power of peers directs SOEs behaviours.

He et al (2012) Dividends behaviour in state- versus family-controlled firms:

Journal of Business Ethics SOEs pay higher and more stable dividends than family companies.

Hung et al Political considerations in the decision of Chinese SOEs to list in Journal of Accounting and SOEs with strong political connections are more likely to list overseas.

(2012) Hong Kong Economics Hou and Moore | Player and referee roles held jointly: The effect of state ownership on | Journal of Business Ethics State ownership amplifies the agency problem Higher state ownership reduces (2011) China’s regulatory enforcement against fraud execution.

Temasek Holdings, a successful SOH ccccccccccccccccceeessesessssssceeeeeeeeeeeeees 37

The success of Temasek Holdings has received the attention of scholars for SOH model (Chen, 2016; Kim & Chung, 2018) Temasek has ultimate ownership in Singapore’s major industries (Chen, 2013) Temasek was set up to “contribute to Singapore’s economic growth by nurturing world-class companies through effective stewardship and commercially driven strategic investments” (Source: www.temasek.com.sg) The statement delivers a commitment to a high level of corporate governance (Ang and Ding, 2006).

Empirical evidences show that Singaporean GLCs have higher valuations and better corporate governance than a control group of non-GLCs (Ang and Ding, 2006).

Companies in which Temasek has direct ownership have more independent directors and are more likely to have an independent director serving as chairman and a higher quality of corporate governance (Chen, 2013) In summary, Temasek’s success is achieved by maintaining high standards of corporate governance (Sam, 2008), rescue from the burden of pursuing social goals and government intervention (Kim & Chung,2018) and autonomy of subsidiaries (Kim & Chung, 2018) However, Temasek model could work properly only in a system where good and clean governance exist (Chen,2013).

Khazanah Nasional Berhad, SOH of Malaysia 555SS++<<<<<<<<s 38

Khazanah was established in 1993 as an state holding company wholly owned by the Government of Malaysia and directly report to the prime minister (chairman of the board) Khazanah has ability to make its investments and activities decisions on purely commercial considerations (Lai, 2012).

Along with GLCs transformation in 2004, Khazanah has responsibility to improve corporate governance in Malaysian GLCs (Lai, 2012) Khazanah has autonomy in directors appointments (Kim & Chung, 2018) Temasek in Singapore and Khazanah inMalaysia have established a reputation for their GLCs management responsibilities(Basu, 2005) Empirical evidences showed that Khazanah ownership positively related to firm GLCs value (Lau & Tong, 2008; Taufil-Mohd et a., 2013).

SCIC, SOH of Viefnam cceeeccccccccesseccccceeeessscccsseeessesccssseeeesesseess 38

SCIC is a SOH (Nguyen et al., 2012) SCIC represent the state capital interests in state-invested companies and become a strategic investor of the government that is capable of generating maximum value and sustainable returns on investments Their missions are to be the government’s strategic investor, active shareholder and a professional financial consultant (www.scic.vn) In 2014, SCIC’s profit was VND 6,900 Billion and it was on top of best performance of SOEs In 2015, revenue of SCIC increased 150% to VND 10,532 Billion and profit also increased up to VND 8,600 Billion basing on investment trading activities (Thanh Thuy, 2016).

Table 2.6 Comparison between Temasek Holdings and SCIC

Temasek Khazanah SCIC History | Incorporated in 1974 † Incorporated in 1993 Incorporated in 2005 Role Manages an Strategic investment Invests and Trades investment fund on fund of the Government | state’s capital and behalf of the of Malaysia managed by Government of Government Singapore.

Mission | Temasek is an active | Utilising a proactive SCIC is investor and | investment approach the Government’s shareholder Undertaking catalytic strategic investor Temasek is a forward- | investments that SCIC is anactive looking institution strategically boost the shareholder

Temasek is a trusted | country's economy SCIC is a professional steward To be the leading | financial consultant regional strategic investment house

BOD BOD is approved by BOD is appointed by the |BOD is approved by

The President Prime Minister The Prime Minister Note: Author

SCIC is one of initiatives of Vietnam’s Government in SOEs reform (Nguyen et al., 2012) The creation of the SCIC is believed as a right step to separate the ownership and regulatory roles of state bureaus to prevent political interference while accumulating the privatization process (VCCI, 2007; Nguyen et al, 2012) SCIC represents Government in equitized SOEs and is expected to mitigate owner-regulator conflicts and increase transparency regarding state’s resources allocation Acting as a controlling stakeholder, SCIC must be atomic investor due to accountabilities for invested companies’ performance Any failure would destroy the intention of the SCIC for the better management of State’s resources In order to achieve this goal, SCIC must aggressively involve into invested company to enhance its value (VCCI, 2007).

SCIC would conduct investments in order to safeguard and develop state capital, use state capital efficiently and improve the operational capacity and competitiveness of subsidiaries (Nguyen et al., 2012)

Table 2.7 Comparison between SCIC and State capital companies and GLCs

State capital companies and GLCs SCIC Objectives Pursue commercial and non-|SCIC pursues the goal of commercial objectives returns on state investment Agency Double agency — concerned about | and adhering to corporate issues self-interested behavior by managers | governance standards and politicians/bureaucrats Transparency | Low level of disclosure Note: Author

SCIC is in line with recently global trend in which government retains ownership and manages SOEs through an intermediate entity (World Bank, 2014) and compatible with OECD principles for SOES (JICA, 2015) SCIC has two main functions of

"managing" and "investing" state capital SCIC has spent most of its financial and personnel resources on state capital management, while new investments are in the early stages, in fact it is a state capital manager (JICA, 2015).

Managing state capital State agency in charge of managing state capital in equitized SOEs Mission | SCIC has no discretion over which SOES to accept

Exercise shareholder rights with market-based considerations.

Divest state capital based on the procedure Objective

The revolution of Temasek is a reference to the development of the SCIC which is now in 2TM stage of “received former state boards” after 1" stage of “started with state capital in start-ups and JVs” SCIC is adopting Temasek’s approaches in which SCIC is administering state capital based on economic considerations and modern corporate governance practice (hands-off approach) as well as establishing internal controls as an independent economic entity (JICA, 2015).

Basically, SCIC has been well performing its role of managing state capital by providing to its invested-companies an autonomy and market-based guidance, a tight controls and financial support for businesses in difficult situations and establishing an effective internal control and organizational structure for equitized SOEs (JICA, 2015).

Following characteristics of SOH defined by Kim & Chung (2018), the features of SCIC are determined through regulations of Decision 151/2005/QD-TTg dated 20/06/2005, Decree 151/2013/ND-CP dated 01/11/2013 and Decree 57/2014/ND-CP dated 16/06/2014 by Government, described in below table:

Table 2.9 SCIC as a Shadow Investor

Relationship | Characteristics | Shadow investor SCIC Government | Fiscal Soft budget constraint (hard | To take responsibility for debts and other asset liabilities

— SOH Management budget constraint on the | of the companies within the limits of charter capital of the surface), discretionary resource transfer companies, to determine and separate assets of the owner from those of the companies (Article 21, Decree 151;

Article 10, Decree 57) Reduction in input controls

Mixed boards of public officials and some outsiders

SCIC BOD has seven (07) members, including full-time members and part-time members The term of office of members of the Members’ Council shall not exceed five (05) years and may be reappointed (Article 26, Decree 57) In the case of officials, public servants and leaders in the State apparatus or political organizations or socio- political organizations, they must be nominated by competent authorities and conformable with the provisions of law (Article 28, Decree 57)

Likely guaranteed SOH autonomy for appointing SOES managers but occasionally violated

To decide on the organizational structures of company management, appoint, re-appoint, relieve from duty,approve the resignation of, sign or terminate contracts with, commend, reward, and discipline chairpersons and members of the Members’ Councils or company presidents, controllers and directors general (directors) of the companies (Article 20, Decree 151; Article 52, Decree 57)

Separation of the role of | Separation of the role of chairman and CEO (Chapter 4, chairman and CEO Decree 57)

SOH-SOEs | Increase in | Performance measures exist | Pilot mechanism of raising responsibility of result controls | between performance and representatives of the Corporation (Article 28, Decree pay; but some gaps existin | 151) practice Rights, obligations, responsibilities and relationship between members and general directors (Section 3, Decree 57)

Note: Author adopted from Kim & Chung (2018)

According to Kim & Chung (2018), SOHs are classified into three groups including corporate investor, shadow investors and submissive investor basing on legal framework and government intervention levels (Table 2.5) Classifications depend on three characteristics including fiscal management, reduction in input controls and increase in result controls SCIC could be classified as a Shadow Investor in according to SOHs characteristics regarding SCIC has hard budget constraint, mixed board members, autonomy in the designation of SOES managers, and a performance measurement mechanism.

While the SCIC has not been able to achieve “Corporate Investor” rating like Temasek Holdings, which has proven to be effective in improving business performance, the

‘Shadow Investor” rating of SCIC indicates that SCIC model could be effective given Malaysian model was demonstrated as improving firm value (Lau & Tong, 2008).

Research Hypothesis Deveẽopmenf << ô<< + + + + ++ssssssssssseeessa 62

Porter (1990) and Abdallah & Ismail (2016) believe corporate governance and ownership structure most strongly determine corporate goals In general, corporate ownership includes five types: Individual, Institutional, Corporate, State and Family (Gedaljovik and Shapiro, 1998) Bhagat & Jefferis (2003) and Iqbal et al (2018) found that corporate performance and ownership structure are interrelated Gedaljovik &

Shapiro (1998) and also found that ownership concentration and corporate profitability are correlated but differs across countries determined by national system of corporate governance This argument is supported in studies of Wang & Shailer (2017) and Yasser & Mamun (2017).

There is a problem of inefficient performance of SOEs Many studies have found that state ownership is often linked to low efficiency (Bai et al., 2004; Ding et al., 2007; Li et al., 2018a; Long & He, 2018) This could be a result of objectives conflicting between macro-economic benefit and shareholder value Nee et al (2007) found that state ownership not only fails to improve firms’ financial performance, but actually impacts negatively on various important firm decisions including personnel, strategic and financial decisions Vietnam started the privatization process in 1992 with limited results (Vu, 2006; Phan, 2015; Nguyen & Tran, 2017) SOEs, on the other hand,receive government subsidies so they do not need to improve their performanceTaussig et al (2015) Therefore, there is reason to believe that state capital firms will be less efficient than other counterparts.

However, Vietnam has The State Capital Investment Corporation (SCIC) which is more specific than other state capital firms SCIC participate into SOEs reforms that are aimed at enhancing the efficiency of state capital utilization Their missions are to be the government’s strategic investor, active shareholder and a professional financial consultant SCIC is SOH who is a strategic investor with profit maximization orientation (Sam, 2010; Kim & Chung, 2018) SOH is more likely to act as an active investor and push for more transparency and better corporate governance to earn long- term profits and as a result, agency problem due to conflict of interests could be overcome (Chen, 2013; Sam, 2013; Kim & Chung, 2018) Besides, SOH as institutional investor who has large enough stake to collect information and monitor the management to focus on corporate performance (Sam, 2013) SOH has less pressure to maximize short term interest to commit for a long-term profit hence it could align the interest of majority shareholder (SOH) and the minority shareholders in SLCs It alleviates the principal-principal agency problem Besides, holding structure is also believed to be able to serve as a layer shielding the SLCs from politics and government intervention (Wicaksono, 2008; Sam, 2010) and reduce the conflict inherent in the state’s roles as both shareholder and regulator (Sam, 2010; Chen, 2013) SOH structure seems to well serve the purpose of resolving the first two problems at state capital firms of conflicted objectives and political interference (Chen, 2013; Nurgozhayeva, 2017) As a result, there is expectation of positive relationship between SOH ownership and firm performance Besides, empirical evidence shows that Singaporean government-linked companies (GLCs) have higher valuations and better corporate governance than a non-GLCs (Ang and Ding, 2006; Kim & Chung, 2018) Besides, empirical evidence in Malaysia showed effectiveness of Malaysian GLCs in creating firm value (Lau & Tong, 2008) Therefore, SOH is expected to have positive impact on firm performance.

Hypothesis 1: SOH ownership has a positive impact on firm performance

Family ownership is popular around the world (Schickinger et al., 2018) In US,Anderson and Reeb (2003) found that family ownership presents in one-third of theS&P 500 firms Many literatures also investigate into the effects of family-control

(Kaserer & Moldenhauer, 2005; Wang & Shailer, 2017) Family ownership could be considered as insider ownership as family businesses traditionally attracted a lot of attention given their predominant economic role (Kaserer & Moldenhauer, 2005;

Wang & Shailer, 2017) Claessens and Fan (2002) argue that family ownership has dominant role in Asia as the concentration ownership in hands of family would lower transaction costs among family members Besides, the problem of information asymmetry between insiders and outsiders could be reduced in family model.

Moreover, hold-up problems, which may prevail in unaffiliated companies, tend to be fewer in family companies (Claessens & Fan, 2002) Wang & Shailer (2017) argue that family ownership generally can enhance monitoring of managers Anderson and Reeb (2003) even found that in U.S family firms outperform non-family firms, thus suggesting that family ownership is an effective organizational structure Yeh et al.

(2001) also found that family representation on the board leads to centralization in authority and decision-making power However, family ownership concentration could increase the expropriation of non-family minority shareholders (Bloom and Van Reenen, 2006) In family companies, unqualified members could be appointed to key positions without competition (Claessens et al., 2000) Moreover, because of close relations and informal linkages, family managers are less to be monitored (Young et al, 2008) Connelly at al (2008) found a significantly lower of firm performance for firms with high family control and family management in Thailand Giovannini (2010) found that in Italy family involvement negatively impacts share performance Wang &

Shailer (2017) found that family ownership and performance change over time and across countries With above arguments, family ownership could be an important element of ownership structure

Hypothesis 2: Family ownership has a negative impact on firm performance.

The flows of foreign investment are important sources of finance for developing countries’ corporates Foreign ownership generally confers performance advantages(Carney et al., 2018) Companies with foreign ownership could achieve benefits from tangible and intangible assets internally transferred to provide a performance advantage to those companies in the market (Dunning, 1988) Foreign institutions enhance capital distribution via investments and innovations (Bena et al., 2017) as well as improve monitoring and corporate governance (Huang &Zhu, 2015) Foreign companies transfer advanced technologies and provide access to international capital markets (Caves, 1996 in Aitken and Harrision, 1999) Pfaffermayr and Bellak (2000) argue that affiliating with foreign firms help local companies have access to newer and superior technologies and lead to superior performance Foreign-owned firms display superior performance in many economies (Gugler, 1998; Bellak, 2004; Estrin et al., 2009; Carney et al., 2018) although Globerman and Shapiro (1999) conclude that there is no significant difference between the productivity of domestic and foreign-owned establishments in Canada This study would like to investigate the role of foreign ownership on firm performance with following hypothesis:

Hypothesis 3: Foreign Ownership has a positive impact on firm performance

Most prior corporate governance studies attempt to find a relationship between boards of directors and firm performance (Brennan, 2006; Pillai et al., 2017; Paniagua et al., 2018) Board of Directors (BOD) should act in the best interest of the company and the shareholders to achieve adequate return and prevent conflicts (OECD, 2004) Agency theorists believes that in order to protect the interests of shareholders, the board of directors must be effective oversight function (Fama & Jensen, 1983; Brennan, 2006;

Paniagua et al., 2018; Masulis et al., 2018) However, the controlling shareholders could affect to company’s strategy through directors of BOD To prevent self- interested actions of BOD that could negatively affect to minority shareholders Codes of corporate governance in many countries including Vietnam regulate the portions of (1) executive directors, (2) non-executive directors as a prevention mechanism.

One of the vital roles of BOD is independence The independence is to provide defense against the exploitative behavior by the controlling shareholders and other directors Independent director is a mechanism to enhance the independence of BOD (Fama & Jensen, 1983; Masulis et al 2018) However, different definitions of

“independence” have been regulated by corporate governance codes around the world(Hu et al., 2009) Mallin (2007, in Hu et al., 2009) defined as having “no relationships or circumstances which could affect the director’s judgment” Independent directors are expected to be active and effective monitoring role than executive (inside) directors (Fama and Jensen, 1983) In case insiders dominate the BOD, instead of focusing on independent directors, outside directors are used as another benchmark for measuring board independence (Johnson et al., 1996) Outside directors are more aggressive in challenging board decisions and act as a counterweight to inside directors (Johnson et al., 1996) Moreover, they could carry external knowledge to firm The independence role allows outsider directors provide advice and resources in helping the firm to succeed (Hillman and Dalziel, 2003) There is a positive association between outside directors and corporate financial performance (Uadiale, 2010) Pearce & Patel (2017) found that board independence is not related to firm performance while Ilhan-Nas at al (2018) found a negative relationship.

In Vietnam, Code of Corporate Governance regulates that one-third of BOD must be non-executive directors who are not CEO, Deputy CEO, CFO or executive officers designated by BOD They could be considered as the balance mechanism between controlling shareholders and minority shareholders Although the Code gives the definition of independence directors with stricter rules, there is not any provision regulates the role of them The impact of independence directors is investigated with following hypotheses:

Hypothesis 4: The proportion of independent directors in the board has a positive impact on firm performance.

The effect of CEO duality on company performance has been widely debated in academic (Dalton at al., 2007; Tang, 2017) Duality grants power authority to one person and this allows CEO control information from other members as well as places the board under control of a manager which reduce the monitoring role of the board(Peng et al., 2007; Desender, 2009; Tang, 2017) Jensen (1993) argues that the BOD is often ineffective because the role of chairperson is combined with CEO position The board is unable to unable to effectively monitor and evaluate the CEO if CEO is also the Chairman (Peng et al., 2007) However, the separation of two roles has both costs and benefits as there could be an implicit rivalry between two roles as well as it is difficult to isolate responsibility for poor performance (Balabat et al., 2004) or even the duality could enhance the unity of commands from BOD (Dalton et al., 2007) The empirical evidences show complicated results (Vo and Nguyen, 2014) Tang (2017) and Li & Patel (2018) found that CEO duality actually negatively impact firm performance Therefore, here it is hypothesizing that the duality of chairman and CEO position may impact on firm performance The sign of impact is hypothesized to be positive as the separation between two roles is expected to prevent the abuse of power.

Moreover, the Codes of U.K and Australia formalize the duality leadership as mandatory.

Hypothesis 5: The duality has a negative impact on firm performance.

3.3.2 Role of Ownership Structure on Firm Performance through impacts of Cash

METHODOLOGY .- G 011111211111 111911 1111881111118 111 re 69 mm ớ aổôdiaiadti

Research MOdelL 1101129101111 1111001111 ng nen 73

ownership structure and board characteristics are examined in the direct relationship with firm performance The indirect relationship is examined between ownership structure and firm performance through cash holdings.

Board Own Characteristics ershi Ownership

Performance cture Cash and Holdings Boar d Characteristics are expected to have direct impacts on firm’s performance with following formula: tirm s performance = f(Ownership Structure, Board Characteristics, Control Variables)

Firm’s performance is dependent variable represented by market value as in Table 4.4.

Ownership Structure is measured by 4 independent variables and Board Characteristics is measured by 2 independent variables as presented in Figure 4.2.

Regression models to determine the nature of the direct relation between Ownership Structure, Board Characteristics and firm performance:

Performance; = Bo + Yi B:Ownership Structure; ae P:Board Characteristics, + io B Control Variables, + fị HOSH Dummy + Yiku12 Bndustry, +

Performance, — Bo + B.Dominant Ownership, + Yj BOwnership Structure Dummies,

+3/_-sOwnership Structure Dummies, x Dominant Ownership; + ya B,Control Variables, + PjHOSE Dummy + Èÿ-ịqPulmdustrn, +

Xp=m+1 B,Year, + e„ (2) where Performance; is the dependent variable representing firm performance (Tobin’s Q

Foreign Ownership Board Independence Duality

Growth rate Leverage Size Board Size Exchange Industry

Figure 4.2 Direct Relationship Model with Variables Author and Market-to-Book) of firm 7 in year ¢, a is an intercept, and e„ 1s the error term “m and are the numbers of industry and year, respectively ỉ is coefficient of these}? relationship between independent variables of the ownership structure and firm performance The sign of this coefficient would be used to test the research hypotheses.

Regression model (1) is demonstrated in details by model (1A) as below:

{TOBIN |MB}y, = Bo + BiSCIC Ownership, + f2Government Ownership; + ;Family

Ownership; + B,Foreign Ownership; + BsBoard Independence; + Bsluality;, + BGrowth Rate, + 6sLeverage, + foSize; +B Board Size, + Bị HOSH Dummy

+ Xk=la Bilndustry, + Xp=m+1 B pYear, + & (LA)

Regression model (2) is demonstrated in details by model (2A) as below:

WẬOBIN | MB}, = Bo + BiDominant Ownership, + B2SCIC Ownership Dummy; + P;SC1C

Ownership Dummy; x Dominant Ownership, + fsGovernment Ownership Dummy; + fBsGovernment Ownership Dummy; x Dominant Ownershipy + Bl amily Ownership Dummy, + ;Family Ownership Dummy, x Dominant

Ownership, + BsGrowth Rate, + PoLeveragey +P 1oSizey + BịHOSH Dummy + yr 12 Bilndustry, + Xp=m+1 B,Yeœ, + & (2A)

Following research models as well as hypotheses development, the regression formula contains 6 independent variables, 4 control variables, stock exchange, industry and year Independent variables include SCIC ownership, government-linked companies’ ownership, family ownership, foreign ownership, board impendence and duality.

Control variables are growth rate, leverage, size and board size fy is intercept and f, (x from 7-7) is the slope while ¢; is error term There are two dependent variables presenting for company performance including Tobin’s Q and Market Book Ratio.

The symbol of “|” is presenting for alternative regression models Due to high correlation with foreign ownership, institutional ownership is excluded from regression model However, this type of ownership is also examined with results are demonstrated in Appendix 3.

There are 6 hypotheses in accordance to research model summarized in the below table.

Table 4.2 Hypotheses for IGMs on Firm Performance

Hypothesis Variables TP ampacts on

H1 SCIC Ownership + H2 Family Ownership - H3 Foreign Ownership + H4 Board Independence + HS Duality - Note: Author

In order to make the research more robust, this study examines the indirect relationship between ownership structure and performance via corporate cash holdings followed by three regression models.

Model Opler et al (1999) is replicated to explore the relationship between ownership structure and corporate cash holdings with ownership structure factors are supplemented basing on model of Ku et al (2013) Model of Opler et al (1999) examined determinants of cash holding at optimal levels which is adopted by many studies on cash holdings such as Ferreira and Vilela (2004), Dittmar and Mahrt-Smith (2007), Bates et al (2009), Schauten et al (2011), Megginson et al (2014), Seifert &

Gonenc (2018) and Loncan (2018) This study contributes Vietnamese ownership structure characteristics into previous studies to consolidate the findings.

Regression model to determine the nature of the relation between ownership structure and corporate cash holdings:

Cash to Net Assets;, = Bo + M ¡ B Capital Structure to Net Assets, + Yi}, B Ownership Structure, + Vr13 Bdndustry, + Yb=m +1 pYear, + & (3) where Cash; is the dependent variable representing cash holdings to net asset of firm 7 in year ¢, a is an intercept, and ¢; is the error term “m” and are the numbers ofse}? industry and year, respectively 6 is coefficient of the relationship between independent variables of the capital structure as well as ownership structure and corporate cash holdings The sign of this coefficient would be used to test the research hypotheses.

Regression model (3) is described in details as below:

Cash to Net Assets;, = Bo + BiMarket to Book to Net Assets;;+ B2Size;, + BsCash Flow to Net Assets;, + 6undustry Sigma,, + BsNet Operating Working Capital to Net Assets,,+ BeCapital Expenditure to Net Assets,, + [Leverage ;, + PsDividend Dummy,, BạSCIC Ownership, + government Ownership, + ,tamiy Ownership, + Bi oreign Ownership, +Pj a4 P; § Pj Me 13 Bdndustry, + Xp=m +1,pYear, + e„(3A)

Details of independent and dependent variables are described in Table 4.9.

There are 4 hypotheses in accordance to impact of ownership structure on cash holdings model summarized in the below table.

Table 4.3 Hypotheses for Impacts on Cash HoldingsHypothesis Variables Expected impacts on

Heyl SCIC Ownership + Note: Author

Following Dittmar and Mahrt-Smith (2007), the investigation on impacts of corporate governance to firm value by using excess cash is explored The same model is explored in studies of Schauten et al (2011), Ku et al (2013) and Seifert & Gonenc (2018) By supplement ownership structure into model of Dittmar and Mahrt-Smith (2007), this study considers the value effects of different types of ownership in term of interaction with level of excess cash on firm value.

Regression model to determine the nature of the relation between ownership structure and firm value in term of interaction with level of excess cash:

Firm Value;, = Bo + Min Capital Structure to Net Assets, + ¡:Ownership, + f¡„FXcess Cash, + B)s;Ownership;x Excess Cash, + Yik Bdndustry, + Xp=m+1 pŸear, + & (4) where Firm Vaiue, is the dependent variable representing firm value (Market Value to Net Assets) of firm 7 in year ớ, a is an intercept, and ứ„ 1s the error term “m” and “I” are the numbers of industry and year, respectively ỉ is coefficient of the relationship between independent variables of the ownership structure and firm performance The sign of this coefficient would be used to test the research hypotheses.

Regression model (4) is described in details as below:

MV; Ej dE; dE; Dj aD; aD; lj dl;jt jt jt }t+2 Jt Jt }t+2 Jt Jt

NAje , 'NA je "NAjp 1? NAje ' ÍNAj NAjt NAje NAjt SNA je! dlj đNA; đNA; dMV; jt+2 jt jt+2 j,t+2

+B | ——— — TT + ị(Wwnership, + fijEXcess Cash, +PNAjp J NAj Pu NAje 12 NAjt B13 Pi ` Pr it

BisOwnership;x Excess Cash, + 3 ệ—1a B dndustry;, + Xp=m +1 pŸear, + & (4A) Details of independent and dependent variables are described in Table 4.10.

Additionally, following Faulkender and Wang (2006), Dittmar & Mahrt-Smith (2007),Schauten et al (2011), Ku et al (2013) and Seifert & Gonenc (2018), the model on value of changes in excess cash is replicated with the purpose of examination on how a change in cash holdings leads to a change in the market valuation of a company with impact of ownership structure The model, however, is different from previous studies in which the ownership structure is supplementary into model of Faulkender and Wang (2006) to explore on the relation between four ownership types on the value of changes in excess cash The change in firm value is measured by the excess return applying Fama & French (1993) size and book-to-market portfolios Same as Dittmar

& Mahrt-Smith (2007), the changes are controlled by firms’ profitability, financial policy and investment policy regarding argument that firm returns are impacted by idiosyncratic characteristics (Dittmar & Mahrt-Smith, 2007).

The dependent variable is the excess return and the independent variable is change in cash, both by itself and interacted with ownership structure The effect of ownership structure is determined by interacting the change in cash with each type of ownership.

It is expected that the change of cash would positively impact company’s market value by controlling for changes in profitability, investment and financing (Dittmar &

Mahrt-Smith, 2007) The ownership structure is also expected to have impact on this change or it 1s expected a positive coefficient on the interaction between cash and owership structure.

The regression equation is described as following formula:

Excess Return, = Bo + an :Capital Structure, + BQOwnership; + ịi Ownership; x ACashy

+ Yik Bidndustry, + Xp=m+1 B,Year, + & (5) where Excess Return, 1s the dependent variable of firm 7 in year ¢, a 1s an intercept, and

& 1S the error term “m” and “ẽ” are the numbers of industry and year, respectively / isse}? coefficient of the relationship between independent variables of the ownership structure and firm performance The sign of this coefficient would be used to test the research hypotheses In details, model 5 is described as below:

AE jt ANA jt + Al jt AC it

Excess Return;, = By + _— — + B> + p3 4, B B Mjt-1 p Mj,t-1 Mj t-1 M jt-4 + Bs + Bo + BolM j,t-1 M jt-1 AD ; Cj

NF jt Cjt—1 AC jt | ủs ‘Bo — + BiOwnershipy + B11 Ownership, x ——+

;k BulndlustrV, + Xp=m+1 B„Ycar, " ej (SA)

Details of independent and dependent variables are described in Table 4.11 There are 4 hypotheses in accordance to impacts of ownership structure on shareholders’ value of cash holdings model summarized in the below table.

Table 4.4 Hypotheses for Impacts on Shareholders’ Value of Cash Holdings

Hsnv1 SCIC Ownership +Note: Author

Research Variables for Impacts of Ownership Structure and Board

Characteristics on Firm Performance 4.2.2.1 Independent Variables

Independent variables of study include S Own, GOwn, F Own, Fr Own, B Indep, D_Dual representing for SCIC ownership, Government Ownership, Family ownership, Foreign Ownership, Board Independence, Duality respectively The independent variable of Dom represents for Dominance Ownership which is ownership of Largest Shareholder minus Board Opponent Ownership Details of measurement of each variable is described in the below table.

Table 4.5 Independent Variables for Impacts of Ownership Structure and Board

Characteristics on Firm Performance Variable | Variable Measurement Previous Studies | Hypothesis

Name Code Multivariate Linear Regression SCIC 5 Own | The percentage of | Author with | HI Ownership company shares owned by | heredity from Ang

SCIC, and Ding (2006) Government | GOwn | The percentage of | Bai et al (2004), Ownership company shares owned by | Ding et al (2007), government (exclude | Cheung et al.

Chizema (2011),Tran and Duong(2011), Phung andHoang (2013), Tran et al (2015), Phung and Mishra (2016),Pillai et al (2017)Li et al (2018a)Family F Own | The percentage of | Isakov and | H2Ownership company shares owned by | Weisskopf (2009), a family To be considered a family firm an individual or a family must be the largest shareholder and hold at least 20% of ultimate voting rights (La Porta et al., 1999).

Fr Own The percentage of company shares owned by foreign investors

Dwivedi and Jain (2005), Phung and Hoang (2013), Ang and Ding (2006), Chen at al (2006), Abdallah & Ismail (2016), Phung &

B_Indep The proportion of independent directors on the board of directors

Kang and Shivdasani (1995), Cheung et al.

Vo and Nguyen (2014), Duong and Vo (2016), Tang (2016) Nguyen et al (2017), Pearce &

Duality D_Dual Dummy variable: 1 if the company has the combination of the roles of Chairman and CEO.

Bhagat and Bolton (2008), Cheung et al (2010), Phan (2013), Vo and Phan (2013), Vo and Nguyen (2014), Duong & Vo (2016), Tang (2016), Pillai et al.

Dom Dominant ownership is the percentage stock ownership of the largest shareholder minus the opponent ownership

Author with heredity fromChen et al (2006)

Dummy D S Do | Dummy variable: 1 if the | Author with SOH m SCIC is the largest | heredity from Ownership shareholder of the | Chen et al (2006) Dominance company

Dummy D_G_Do | Dummy variable: 1 ifthe | Author with Government m government is the largest | heredity from Ownership shareholder of the Chen et al (2006) Dominance company

Dummy D_Famil | Dummy variable: 1 ifthe | Author with Family y family is the largest heredity from Ownership shareholder of the Chen et al (2006) Dominance company

SCIC ownership is most important factor in this model regarding SOH is new model not only in Vietnam but also around the world Previous studies in Vietnam considered state ownership is a unified entity regardless there are different types of state ownership SOH is repented for new era of state capital firms corporate governance with its effectiveness demonstrated in Singapore This study would provide the examination on SOH model in weak corporate governance environment.

Government ownership, which was explored by previous studies in Vietnam, continues being explored by this study However, with exclusion ownership of SCIC, government ownership could provide a sharper picture on its ultimate role.

Family ownership is not a new object of global research However, in Vietnamese context, this variable is not explored many studies in corporate governance Therefore, the finding on impact of family ownership could contribute understanding of its role in Vietnam and consolidate for agency theory on principle-principle conflicts.

Next, foreign ownership is demonstrated as effectiveness in previous studies The finding on its impact in this study could consolidate the critical role of foreign ownership in Vietnamese ownership structure.

Dominant ownership, which is measured by the percentage stock ownership of the largest shareholder minus the opponent ownership, is used to demonstrate for the argument of different types of dominant shareholders have different objectives and these goals will affect firm performance However, it is different from original study of Chen et al (2006), Gan et al (2017) in which dominant ownership is measured by largest ownership minus second largest ownership In this study, opponent ownership is using instead of second largest ownership with argument that the ownership of members other than largest shareholder could act as constricted entity to the role of largest shareholder If ownership of other members is large enough, it could resist the dominance of largest shareholder.

The institutional ownership, which is found to have positive effects on firm performance as well as enhance corporate governance in many studies (Balatbat et al., 2004; Chen et al., 2008), is not offered in this research model due to correlation with foreign ownership (>0.5) However, the model with institutional ownership is presented in Appendix to demonstrate for robustness results of SOH ownership.

According to Richard et al (2009), firm performance should be measured appropriately for study purposes and until now, financial performance is a dominant measurement for corporate performance (Richard et al., 2009) Krafft et al (2013) indicated that there are three types of performance often used in studies of corporate governance including investment performance, market performance and operating performance This study would use market performance measurement to investigate the impacts of ownership structure on firm performance.

Dependent variables include Tobin and MB for Tobin’s Q and market to book value ratio.

Table 4.6 Performance Measurements for Impacts of Ownership Structure and Board

Characteristics Measurement Variable Code | Definitions Market Performance

Tobin’s Q Tobin Market value of equity plus book value of total liabilities divided by Total Assets Market to Book MB Market value of equity divide by Book

Value of Equity Note: Author compiled

Tobin's Q ratio is proposed by James Tobin in 1969 The idea of Tobin’s Q is that if the market assesses a company’s value higher than the value of physical assets, it is the signal that this company has prospects for development If Q is greater than 1, it shows that the market highly values the firm’s assets and growth opportunities Also, the high ratio indicates that the investors anticipate management could create more value from current assets If the corporate governance could improve market valuation of the company, this should translate into a higher firm valuation.

The Market to Book ratio (MB) is used to assess a company’s current market value relative to its book value The market value is the current stock price of all outstanding shares while the book value is price of outstanding shares on balance sheet This ratio would express the market’s perception of a specific stock value and shows how much an investor is paying for each dollar in net asset of that company.

Besides the impacts of vary dependent variables, performance of listed companies, could be affected by a number of environmental variables inside and outside of the company Based on literature review, this study includes Growth rate, Leverage, Firm Size, Board Size, Stock Exchange, Industry and Year as control variables with details in below table.

Table 4.7 Control Variables for Impacts of Ownership Structure and Board

Characteristics on Firm Performance Control Variable Measurement Reference Variable Code

Growth Rate Growth | Compounded Klapper and Love (2004), King annualized rate of | and Santor (2008), Tran and growth of revenue | Duong (2011)

Leverage Lev Leverage = (book | Balatbat et al (2004), Tran and(%) value of |Duong (2011), Le and Chizema debt)/(total assets) | (2011), Vo and Phan (2013),

Cheung et al (2014), Limpaphayom et al (2015), Duong & Vo (2016), Tang (2016), Pillai et al (2017), Pearce

& Patel (2017) Firm Size Size The natural | Hu et al (2009), Tran and Duong logarithm of total | (2011), Le and Chizema (2011), assets Cheung et al (2014),

Limpaphayom et al (2015), Tang(2016), Pillai et al (2017), Pearce

(2018) Board Size B Size | The total number | Yermack (1996), Postma et al. of directors on a | (2003), Dwivedi and Jain (2005), board of directors | Hu et al (2009), Vo and Phan

(2013), Phan (2013), Vo and Nguyen (2014), Limpaphayom et al (2015), Duong and Vo (2016), Pearce & Patel (2017)

Stock D Hose | Dummy variable: | Chen at al (2006) Exchange 1 if the company is listed on HOSE Industry Industry | Industry Lemmon and _ Lins (2003),

Classification of | Limpaphayom et al (2015), Pillai ICB et al (2017), Pearce & Patel

(2017) Year Year Dummy variable | Tran and Duong (2011), Vo and for year Phan (2013), Tang (2016) Note: Author compiled

Firm size is often used as control variable in research related firm financial performance The largest enterprise is easy to be recognized and would be received more investment opportunities The large firm expected to have a positive effect on performance The listed Exchange (HOSE) could also effect to firm performance as HOSE requests for stringent criteria than HNX Company performance could be different between different industries because of supply chains or external environment Industries in this study are classified by the Industry Classification Benchmark (ICB) This classification was launched by Dow Jones and FTSE in 2005 including 10 industries of Oil and Gas, Basic Materials, Industrials, Consumer Goods, Health Care, Consumer Services, Telecommunications, Utilities, Financials and Technology.

Table 4.8 ICB Industry ICB Code _ | Industry

1 Basic Materials 2 Consumer Goods 3 Consumer Services 4 Financial

Utilities 0 Health Care Note: Industry Classification Benchmark (2016)

There could be impacts of Year on performance as the period of 2009-2013 is the recession after financial crisis in 2008.

Research Variables for Impacts of Ownership Structure on Firm

4.2.3.1 Variables for Impacts of Ownership Structure on Corporate Cash

Regression model (3A) determines the nature of the relation between ownership structure and corporate cash holdings in which these independent variables are using:

Table 4.9 Variables and Definitions for Cash Holdings Model

Independent Variables Market to Book to | The book value of assets, less the book value of equity, plus Net Assets the market value of equity, divided by Net Assets.

Size The natural logarithm of total assets Cash Flow to Net} Ratio of Cash Flow to Net Assets; Cash Flow = Earnings Assets before interest and tax plus depreciation; Net Assets = Total

Assets — Cash and Cash Equivalent Industry Sigma The mean of standard deviations of cash flow over net assets over 5 years, for firms in the same industry, as defined by ICB Code

Net Operating | Ratio of Net Operating Working Capital to Net Asset; Net Working Capital to | Operating Working Capital = Current Assets — Current Net Assets Liabilities — Cash and Cash Equivalent; Net Assets = Total

Assets — Cash and Cash Equivalent Capital Expenditure | Ratio of Capital expenditure to Net assets Capital to Net Assets Expenditure is Investment Cash Flow.

Leverage Ratio of debt to net assets Dividend Dummy — | Dummy variable of dividend distribution: 1 if indicated dividend have been distributed SCIC Ownership The percentage of company shares owned by SCIC.

Government The percentage of company shares owned by government Ownership (exclude SCIC)

Family Ownership | The percentage of company shares owned by a family To be considered a family firm an individual or a family must be the largest shareholder and hold at least 20% of ultimate voting rights (La Porta et al., 1999).

Foreign Ownership | The percentage of shares owned by foreign investors

Control Variables Industry Industry Classification of ICB Year Dummy variable for year

Cash to Net Assets | In[(Cash)/(Total Assets — Cash and Cash Equivalent) ]

Note: Author compiled As explaining above, regression model (3) is inherited from Opler et al (1999) Model of Opler et al (1999) examined determinants of cash holding at optimal levels which is adopted by many studies on cash holdings such as Ferreira and Vilela (2004), Dittmar and Mahrt-Smith (2007), Bates et al (2009), Schauten et al (2011), Megginson et al (2014), Bruce & Gonenc (2018) This study follows Ku et al (2013) and supplements Vietnamese ownership structure characteristics including SCIC Ownership, Government Ownership, Family Ownership and Foreign Ownership into regression model.

4.2.3.2 Variables for Impacts of Ownership Structure, Cash Holdings and

Continuing, regression model (4A) determines the nature of the relation between ownership structure and firm value in term of interaction with level of excess cash in which these variables are using:

Table 4.10 Variables and Definitions for Level of Excess Cash Model

Independent Variables E/NA Earnings before Interest and Tax to Net Assets D/NA Dividends to Net Assets

I/NA Interest Expense to Net Assets NA/NA Net Assets = Total Assets — Cash and Cash Equivalent MV/NA Firm’s market value to Net Assets

Excess Cash Excess Cash Ratio = the residual value of regression from model (3A) above Ownership One of 4 ownership types including SCIC Ownership,

Government Ownership, Family Ownership and _ Foreign Ownership

Control VariablesIndustry Industry Classification of ICBYear Dummy variable for year

MV/NA | Firm’s market value to Net Assets

Note: Author compiled Model (4A) is inherited from Dittmar and Mahrt-Smith (2007) with Vietnamese ownership structure characteristics supplementary In this regression model, term /-2 is the base period, term ¢ is the current period, and / +2 is the next period đ(variables, ,) is a certain variable’s value at the current period value minus its value from the base period while đ(variables,,.›) is the next period’s value minus current period’s value.

The model controls for past changes, future changes, and current levels of Earnings, Dividends, Interests Expenses to reflect forecast of investors (Ku et al., 2013; Seifert

Regression model (5A) determines the nature of the relation between ownership structure and firm value in term of interaction with change of excess cash in which these variables are using:

Table 4.11 Variables and Definitions for Change of Excess Cash Model

Independent Variables AC/M Chang in Cash to Market Value of Equity AE/M Chang in EBIT to Market Value of Equity ANA/M Chang in Net Assets to Market Value of Equity

AI/M Chang in Interest Expense to Market Value of Equity AD/M Chang in Dividend to Market Value of Equity

C/M Cash to Market Value of Equity

L Leverage = Total Debt/(Total Debt + Market Value of Equity) NF New Finance = Net New Equity Issues + Net New Debt Issues

Ownership One of 4 ownership types including SCIC Ownership,

Government Ownership, Family Ownership and _ Foreign Ownership

Control Variables Industry Industry Classification of ICB Year Dummy variable for year Dependent Variables

Excess Return | Annual excess returns on 25 portfolios formed on size and book-to-market ratio factors adopted Fama and French (1993) methodology.

Where AX indicates a change in X from year / — / to t Mj, = Market Value of Equity at time / C;,„; = Cash at time ¢-/ £;, is Earning before Interest and Tax from year / — / to t NA; is Net Asset at time 7 l„ = Interest Expense from year /— 7 to / Dj = Dividend Payout from year / — / to t L„ = Debt„(Deb„+Mí,) = leverage at time ¢ Defb„ = Short term debt ;, + Long term debt „ NF; = New Finance from year ¢ — / to t = Net New Equity Issues + Net New Debt Issues Ownership;, = Ownership Type.

In Models (3A), 4(A) and (5A), year dummies are included to capture macroeconomic and time trend effects, as well as industry dummies to capture industry effects.

In this study, the ordinary least squares (OLS), Fixed Effects (FE), Random Effects (RE), Feasible Generalized Least Square (FGLS) and Panel-Corrected Standard Error (PCSE) regression techniques are used to test the hypotheses Model 1A and 3A are the regressions for multivariate analysis while 2A and 4A are interaction variables regressions These models are estimated by OLS, FE, RE, FGLS and PCSE Before regression, correlation analysis would be conducted to ensure for non-multicollinearity between the variables used in the model If there are multicollinearity between independent variables, regression analysis can have severe effects on the estimated parameters and on the estimation techniques.

Data using in this study is Panel Data which combines cross-section and time-series data Because all the cross-sectional units have the same number of time series observations so the panel is balanced By combining data in two dimensions, panel data gives more data variation, less collinearity and more degrees of freedom.

Pooled Cross Section (POLS) ceccccecessssccceeeeeeeeeeseeessssnceeeeeeeeeeeeees 89

Pooling cross section would create bigger sample sizes in which the effect of time whether relationships have changed over time are investigated However, the assumptions for POLS are intercept and slope coefficients are constant across time and firms and that the error term captures differences over time and over firms As a result,pooled regression by POLS may result in heterogeneity bias.

Fixed Effects (FE) S111 g0 000 1 1 vn ng vờ 90

FE is using when only analyze the impact of variables that vary over time FE explore the relationship between predictor and outcome variables within an entity Each entity has its own individual characteristics that may or may not influence the predictor variables (Torres-Reyna, 2007) FE controls for impact or bias from individual characteristics FE removes the effect of those time-invariant characteristics and the net effect of the predictors on the outcome variable could be assessed (Torres-Reyna, 2007).

FE assumes that those time-invariant characteristics are unique to the individual and should not be correlated with other individual characteristics Each entity is different therefore the entity’s error term and the constant should not be correlated with the others If the error terms are correlated, then FE is no suitable rather than RE (Torres- Reyna, 2007).

Heteroscedasticity is variance of the error term is not constant Standard errors are biased when heteroskedasticity is presented in model As a result, it leads to statistics and confidence intervals bias (William, 2015).

To detect heteroscedasticity, Breusch-Pagan/Cook-Weisberg Test in linear model is using in which null hypothesis is error variances are all equal This test is applied for OLS estimation (William, 2015) For FE estimation, Wald test is used (Baum, 2001) while Breusch-Pagan Lagrange Multiplier (LM) Test is used (Drukker, 2003).

In the classical linear regression model, it is a simplification to assume that error terms are independent If error terms of observations are correlated, it is autocorrelation (Cocharane & Orcutt, 1949) Autocorrelation is popular in panel-data models and it biases the standard errors and lead the results to be less efficient (Drukker, 2003).

Wooldridge test is one of methodology to test for correlation in random- or fixed- effects model (Drukker, 2003).

RE assumes that the variation across entities is random and uncorrelated with the predictor or independent variables (Torres-Reyna, 2007) Random effects assume that the entity’s error term is not correlated with the predictors which allows for time- invariant variables to play a role as explanatory variables (Torres-Reyna, 2007).

4.3.6 Feasible Generalized Least Square (FGLS)

OLS has disadvantage if the model contains heteroskedasticity and/or autocorrelation even with fixed effect or random effect given the assumption that covariance between independent variables and the error term is zero (Vo and Nguyen, 2014) Feasible Generalized Least Square (FGLS) is more suitable for panel data and has more advantages than pool OLS especially in case of the presence of heteroskedasticity, serial correlation or non-zero covariance between independent variable and error term (Wooldridge, 2002 in Vo and Nguyen, 2014).

4.3.7 Panel-Corrected Standard Error (PCSE)

Feasible Generalized Least Square (FGLS) is more suitable for panel data, however, FGLS assumes that the error process is known but not estimated and could lead to extreme overconfidence or underestimating variability unless there are substantially more time points (T) than there are cross-sectional units (N) (Beck & Katz, 1995).

As a result, time-series cross-section data should be proceed lagged dependent variable or transforming the data to eliminate serial correlation of the errors using PCSEs (Beck

& Katz, 1995) PCSEs are more accurate than the usual OLS standard errors in the presence of panel error structures (Beck & Katz, 1995; Beck, 2001).

4.3.8 Breusch-Pagan Lagrange Multiplier (LM) Test and Hausman Test

This test varifies the appropriateness of RE or OLS regression The null hypothesis in the LM test is that variances across entities is zero meaning no significant difference across units (Torres-Reyna, 2007).

Hausman Test is used to test whether the error term is correlated with the repressors,the null hypothesis is they are not This tests for the statistical significance of the difference between the coefficient estimates obtained by FE and by RE, under then null hypothesis that the RE estimates are efficient and consistent, and FE estimates are inefficient (Torres-Reyna, 2007).

Demsetz (1983) argued that ownership structure should be regarded as the endogenous result of decisions for shareholders profit-maximization goal in which optimal ownership might vary regarding characteristics of the company Demsetz & Villalonga (2001), moreover, argued that insider information of corporate performance wou Id lead to changes in insider ownership structure given managers would change thier ownership based on company performance Holderness (2003) went even further to argue that performance might be the one-way decisive factor in ownership structure given performance rewards to managers are often stock options.

The concern on above arguments, however, still remains regarding ownership structure is stability over time which makes it is more reliable that ownership is exogenous to performance (Gugler & Weigan, 2003) Since Jensen & Meckling (1976), ownership is considered a causal exogenous factor determining the performance of the company (Hu & Izumida, 2008) Morck et al (1988) disagreed with the view of Demsetz (1983) and provided empirical evidence of relationship between ownership structure & firm performance Besides, Himmelberg et al (1999) who argued that endogeneity of ownership is determined by exogenous changes in firm’s environment Himmelberg et al (1999) concluded that shareholders decide ownership structure optimally Gugler &

Weigan (2003) empirically found that large shareholders are exogenous to performance.

Technically, endogeneity problem of ownership and performance could be solved by using panel data and control fixed individual effect (Hu & Izumida, 2008) However, fixed effect might not detect the effect of ownership structure on company performance given stability of ownership over time could not able to lead to substantial change in performance (Zhou, 2001).

Given endogeneity was argued by several researchers such as Demsetz (1983) andDemsetz (1983) although this view was being challeged by Gugler & Weigan (2003), this study takes into consideration the endogeneity problem to ensure the consistency of the results Beside lagged, predicted and residual values could help control for endogeneity (Coles at al., 2006) Instrumental variable techniques (IV) could be use to control for potential endogeneity variables (Hermalin & Weisbach, 1991; Gugler &

Weigan, 2003; Dzi lamé, 2012; Xi & Cheng, 2016) in which lagged explanatory variable is used as instrument for ownership structrure (Hermalin & Weisbach, 1991;

Gugler & Weigan; 2003; Dzi anié, 2012) Besides, lagged values could be used to estimate simulteneity between ownership and performance (Xi & Cheng, 2016).

Chapter 4 described the sampled data and data sources as well as variables definitions.

The methods used to test the hypotheses are also explained in this chapter The research model is presented to provide an intuitive view on the relationship between factors Furthermore, hypotheses development is also summarized with relevant variables proposed in research model.

Autocorrelation T©SfS -c c1 1121 v1 ng ven 90

In the classical linear regression model, it is a simplification to assume that error terms are independent If error terms of observations are correlated, it is autocorrelation (Cocharane & Orcutt, 1949) Autocorrelation is popular in panel-data models and it biases the standard errors and lead the results to be less efficient (Drukker, 2003).

Wooldridge test is one of methodology to test for correlation in random- or fixed- effects model (Drukker, 2003).

Random Effects (RE) -GG 1130001111111 1 111111 ng 91

RE assumes that the variation across entities is random and uncorrelated with the predictor or independent variables (Torres-Reyna, 2007) Random effects assume that the entity’s error term is not correlated with the predictors which allows for time- invariant variables to play a role as explanatory variables (Torres-Reyna, 2007).

Feasible Generalized Least Square (FGLS) 0 cccceeesessceceeeeeeeeeeees 91

OLS has disadvantage if the model contains heteroskedasticity and/or autocorrelation even with fixed effect or random effect given the assumption that covariance between independent variables and the error term is zero (Vo and Nguyen, 2014) FeasibleGeneralized Least Square (FGLS) is more suitable for panel data and has more advantages than pool OLS especially in case of the presence of heteroskedasticity,serial correlation or non-zero covariance between independent variable and error term(Wooldridge, 2002 in Vo and Nguyen, 2014).

Panel-Corrected Standard Error (PCSE) . -cSSsssssesssssss 91

Feasible Generalized Least Square (FGLS) is more suitable for panel data, however, FGLS assumes that the error process is known but not estimated and could lead to extreme overconfidence or underestimating variability unless there are substantially more time points (T) than there are cross-sectional units (N) (Beck & Katz, 1995).

As a result, time-series cross-section data should be proceed lagged dependent variable or transforming the data to eliminate serial correlation of the errors using PCSEs (Beck

& Katz, 1995) PCSEs are more accurate than the usual OLS standard errors in the presence of panel error structures (Beck & Katz, 1995; Beck, 2001).

Breusch-Pagan Lagrange Multiplier (LM) Test and Hausman Test

This test varifies the appropriateness of RE or OLS regression The null hypothesis in the LM test is that variances across entities is zero meaning no significant difference across units (Torres-Reyna, 2007).

Hausman Test is used to test whether the error term is correlated with the repressors,the null hypothesis is they are not This tests for the statistical significance of the difference between the coefficient estimates obtained by FE and by RE, under then null hypothesis that the RE estimates are efficient and consistent, and FE estimates are inefficient (Torres-Reyna, 2007).

Demsetz (1983) argued that ownership structure should be regarded as the endogenous result of decisions for shareholders profit-maximization goal in which optimal ownership might vary regarding characteristics of the company Demsetz & Villalonga (2001), moreover, argued that insider information of corporate performance wou Id lead to changes in insider ownership structure given managers would change thier ownership based on company performance Holderness (2003) went even further to argue that performance might be the one-way decisive factor in ownership structure given performance rewards to managers are often stock options.

The concern on above arguments, however, still remains regarding ownership structure is stability over time which makes it is more reliable that ownership is exogenous to performance (Gugler & Weigan, 2003) Since Jensen & Meckling (1976), ownership is considered a causal exogenous factor determining the performance of the company (Hu & Izumida, 2008) Morck et al (1988) disagreed with the view of Demsetz (1983) and provided empirical evidence of relationship between ownership structure & firm performance Besides, Himmelberg et al (1999) who argued that endogeneity of ownership is determined by exogenous changes in firm’s environment Himmelberg et al (1999) concluded that shareholders decide ownership structure optimally Gugler &

Weigan (2003) empirically found that large shareholders are exogenous to performance.

Technically, endogeneity problem of ownership and performance could be solved by using panel data and control fixed individual effect (Hu & Izumida, 2008) However, fixed effect might not detect the effect of ownership structure on company performance given stability of ownership over time could not able to lead to substantial change in performance (Zhou, 2001).

Given endogeneity was argued by several researchers such as Demsetz (1983) andDemsetz (1983) although this view was being challeged by Gugler & Weigan (2003), this study takes into consideration the endogeneity problem to ensure the consistency of the results Beside lagged, predicted and residual values could help control for endogeneity (Coles at al., 2006) Instrumental variable techniques (IV) could be use to control for potential endogeneity variables (Hermalin & Weisbach, 1991; Gugler &

Weigan, 2003; Dzi lamé, 2012; Xi & Cheng, 2016) in which lagged explanatory variable is used as instrument for ownership structrure (Hermalin & Weisbach, 1991;

Gugler & Weigan; 2003; Dzi anié, 2012) Besides, lagged values could be used to estimate simulteneity between ownership and performance (Xi & Cheng, 2016).

Chapter 4 described the sampled data and data sources as well as variables definitions.

The methods used to test the hypotheses are also explained in this chapter The research model is presented to provide an intuitive view on the relationship between factors Furthermore, hypotheses development is also summarized with relevant variables proposed in research model.

The objective of this chapter presents the results of research, which describes the dataset, correlations among the variables, regression results of the research models and hypotheses Finally, the discussion of results is presented.

To construct the dataset for our study, we include all listed companies on HOSE, the most important stock market in Vietnam, which represents for more than 90% market capitalization of Vietnam stock market Our initial sample consists of entire population of HOSE firms from 2005 to 2017 with no missing data for the main variables used in the analyses Due to the inception of SCIC in 2005, the listing of SCIC-Linked companies only started in 2005, therefore, the period of 2009 to 2017 becomes the most effective period for the study of the impact of SCIC roles in listed companies.

Our final sample includes 242 firms which forms a balanced panel dataset of 2,137 firm-year observations with descriptive statistics being presented in Table 5.1.

Data are collected from the annual reports and prospectuses published on HOSE and HNX along with audited financial statements provided by Tai Viet Corporation (Vietstock), Ho Chi Minh City Securities Corporation (HSC) and VietCapital Securities Joint Stock Company (VCSC) The ownership data were manually obtained from each annual reports and prospectuses These data were verified with transaction recorded by VCCorp Corporation (CafeF) subjecting to compulsory information disclosure, especially for family member’s ownership which are only published under each related parties’ transactions There were 2,137 observations from these companies on testing period forming balanced panel data.

Independent variables of study include S Own, G_ Own, F Own, Fr Own, B Indep,B_Size representing for SCIC ownership, government ownership, family ownership,foreign ownership, board independence and duality Dom is representing for dominant ownership Dependent variables include Tobin and MB representing for Tobin’s Q and market to book value ratio, respectively.

S' Own is SCIC ownership G_Own is Government Ownership # Own 1s Family Ownership Fr Own 1s Foreign Ownership.

Dom is dominant ownership B Jndep is Board Independence D Dual is Duality Size is company size, Lev is Leverage.

Growth is company’s growth BSize is Board Size D_ Hose is listed exchange Tobin and MB are Tobin’s Q and Market to Book ratio Sample includes 242 firms for the period of 2009 to 2017.

Stats N Mean Median Min Max Standard

Deviation sown 2137 0.0312 0 0 0.578 0.101 g own 2137 0.243 0.192 0 0.844 0.243 f own 2137 0.0482 0 0 0.810 0.144 fr own 2137 0.0891 0.0230 0 0.882 0.135 dom 2137 0.308 0.304 0 0.873 0.206 b indep 2137 0.119 0 0 0.833 0.169 d dual 2137 0.282 0 0 1 0.450 SIZe 2137 27.22 27.19 23.15 33.27 1.594 lev 2137 0.512 0.543 0.0056 0.995 0.228 growth 2137 0.145 0.0822 -1.039 30.15 0.822 b size 2137 5.540 5 2 I1 1.160 d hose 2137 0.492 0 0 1 0.500 tobin 2137 1.053 0.935 0.0841 11.35 0.615 mb 2137 1.050 0.799 0.0213 9.851 0.995

Descriptive statistics tables also presented the descriptive statistics of the values of the control variables as Growth rate (Growth), Leverage (Lev), Firm Size (Size), Board Size (B Size), Year (Year) and Stock Exchange listed (D Hose) Industry are classified by the Industry Classification Benchmark (ICB) This classification was launched by Dow Jones and FTSE in 2005 including 10 industries of Oil and Gas, Basic Materials, Industrials, Consumer Goods, Health Care, Consumer Services, Telecommunications, Utilities, Financials and Technology.

Descriptive statistics for independent variables (Table 5.1) shows that SCIC owns average at 3.1% of companies’ shares in which SCIC owns max at 57.8% of shares.

SLCs accounted for 14% of total sample with 34 companies The government excluding SCIC owned 24.3% of companies’ shares on average in which maximum ownership is recorded at 84.4% and it is belonged to Oil&Gas industry which is an industry that the government controls to ensure energy security Government is represented in 179 companies, accounting for 74% of the sample Family owns 4.8% on average which maximum ownership is recorded at 81% However, to be considered as a family company, family members must own at least 20% of company shares The foreign investors own 8.8% of total shares with 88% is highest ownership from a foreign investor in a private company (TTP) after Law of Investment 2014 and Decree 60/2015/NĐ-CP take effect and uplift the limitation on foreign ownership The highest state ownership is recorded at 84% at COM 81% is a highest ownership of a family at ST8 The dominant ownership is recorded at 87% found at LGC The most highest board independent ratio is 83% found at SAV in 2017.

Generally, the foreign investors are limited to ownership of listed companies at the threshold of 49% before 2014 This threshold is not controlled by exchanges trading system Therefore, in some cases, foreign investors buy over the regulated number so that the rate of 49% is violated with variance less then 1% Foreign investors then have to sell to ensure the rate is below the regulated level However, the annual reports in some situations disclosed the actual ownership of foreign investors at the reporting time.

RESULTS - - - E1 1333211181113 1111811 1111881111 118111 ray 94

Performance Comparisons between SLCs, GLCs and non-SLCs 98 5.5 _ Regression Results .ccccccccccccccccccsssssessssssseceeeecccceeeesesesssnseeeeeeeeeeeeseeseeeeeaas 104

The main purpose of this study is to demonstrate the role of SCIC as the professional investors owned by the SOH-Linked companies (SLCs) could impact on firm performance in positive way than other state-owned morphological with previous chapters’ arguments The superior SLCs performance is demonstrated by comparing various measures of financial and market performance of SLCs and Government-Linked Companies (GLCs)/non-GLCs presented in Table 5.4 and Table 5.5 presented at Annex 1 These include profitability, asset efficiency, expenditures and financial ratios comparisons GLCs are Government-Linked Companies excluding SLCs andNon-GLCs is remaining.

Following methodology of Ang and Ding (2006) and Chen et al (2006), this study capture the firm profitability by Tobin’s Q, Market-to-Book ratio, ROE, ROA, ROS, Excess Return, P/E The financial ratios for comparisons includes leverage, growth rate and long term debt to equity Asset efficiency is measured by sales divided by total assets and total assets to equity Expenditures are compared by expense to total assets and expense to total assets Finally, investment is measured by capital expenditures to sales and capital expenditures to assets.

There are two groups of comparisons including SLCs versus GLCs (Panel A) and SLCs versus non-GLCs (Panel B) for same criteria as in Table 5.4 and 5.5 The significances of differences in performances are tested via parametric ¢-statistics.

Sample is divided into 3 groups basing on ownership of major owners with 20% control threshold meaning SLCs are companies owned by SCIC at or above 20% The same applied for GLCs and the last group includes companies which is neither SLCs nor GLC.

Market values based performance measures are examined through Tobin’s Q, market- to-book (MB) ratios, excess return and price-to-earnings per share (P/E) The Tobin’s Q and MB are significantly higher among SLCs than both GLCs and non-GLCs at the 1% & 5% levels The Tobin’s Q is significant across all years for SLCs and GLCs comparisons whilst almost years for SLCs and non-GLCs comparisons MB is significant almost years for both group comparisons This demonstrates that the market pays higher value on SLCs relative to their book value Nonetheless, the mean differences in excess return and P/E ratio between SLCs and both GLCs and non- GLCs are not significant General conclusion could be determined from the evidence is that SLCs either outperform, or at least do not under-perform, in comparisons to both GLCs and non-GLCs in terms of market valuation and supports the argument that SCIC ownership spreads (or does decreases) firm value The result is similar to finding of Ang and Ding (2006).

The financial profitability is examined by ROE and ROA in which SLCs outperformGLCs at 1% significant level for ROA SLCs outperform non-GLCs for ROA and

ROE at 1% significant level This finding is consistent with Ang and Ding (2006) in which SLCs are more profitability or at least equal to GLCs or non-GLCs.

Table 5.4 Benchmarking of SLCs and GLCs on Performance Ratios, Privileges and Business Risk

Panel A 2009 2010 2011 2012 2013 Variables SLCs | GLCs T-Test SLCs | GLCs T-Test SLCs GLCs T-Test SLCs GLCs T-Test SLCs GLCs T-Test Tobin's Q 1.344 1.087 0.257** 1.264 1.056 0.209** 1.063 0.809 0.254*** 1.195 1.011 0.184 1.508 1.054 0.454 Market to Book 1.448 1.085 0.364* 1.478 0.908 | 0.570*** 1.597 0.905 0.692*** 1.505 0.908 0.597% ** 1.532 0.903 0.629% **

Expense to TotalAsses | 0.134 | 0.096 | 003§ | 0174 | 0139 | 0035 0.186 0.143 0.043 0.207 0.154 0.053 0.226 0.173 0.053 Expense to Sales 0.127 | 0099 | 0029 | 0158 | 0128 | 0.029 0.159 0.137 0.022 0.165 0.169 -0.004 0.186 0.185 0.001 Interest to Net Assets 0.013 | 0.013 0 0.021 | 0019 | 0002 0.026 0.028 -0.001 0.029 0.027 0.002 0.017 0.02 -0.003

Interest to EBIT 0.141 | 0172| -0.03 | 0214 | 0226 | -0012 0.288 0.206 0.082 0.39 0.309 0.082 0.2 0.409 -0.208 Dividend to Net Assets | 0033 | 0045 | -0012 | 001 | -0.001 | 0011 0.017 0.003 0.014** 0.004 0.004 0 0.001 -0.003 0.004

Long Term Debt to Equtty | 0.316 0.471 -0.155 0.129 0.448 -0.319 0.137 0.447 -0.31 0.107 0.405 -0.298 0.051 0.325 -0.274 Total Assets to Equity 2.17 2.879 -0.710* 2.32 2.851 -0.531 2.447 2.926 -0.478 1.96 2.92 -0.961** 1.954 2.875 -0.922**

Sales to Total Assets 1.282 1.3 -0.018 1.221 1.322 -0.101 1.249 1.452 -0.203 1.443 1.354 0.089 1.221 1.331 -0.111 CapEx to Total Assets 0.035 0.008 0.027 -0.002 | -0.009 0.007 0.02 -0.004 0.025 0.071 -0.101 0.172 0.109 -0.04 0.149 CapEx to Net Assets 0.04 0.011 0.029 -0.004 | -0.009 0.005 0.022 -0.005 0.026 0.077 -0.112 0.19 0.13 -0.04 0.17

CapEx to Sales 0.021 0.033 -0.013 -0.079 | 0.018 -0.097 0.03 0.009 0.021 0.049 -0.133 0.182 0.074 -0.048 0.121 Cash to Total Assets 0.064 0.06 0.003 0.061 0.043 0.018* 0.059 0.043 0.017 0.046 0.042 0.005 0.061 0.052 0.01

Cash to Net Assets 0.069 0.065 0.004 0.055 0.049 0.006 0.071 0.054 0.017 0.049 0.05 -0.001 0.069 0.071 -0.001 Cash to Sales 0.056 0.07 -0.015 0.044 0.053 -0.009 0.058 0.052 0.006 0.048 0.078 -0.03 0.143 0.068 0.076 EBITDA SD 1.10E+11 | 3.80E+10 | 6.9e+10** | 1.10E+11 | 5.70E+10 | 5.00E+10 | 1.60E+11 | 5.80E+10 | 1.0e+11**

Asterisks (***, ** and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively This table compares the means of various performance measures of SOH-Linked Companies to GLCs The mean difference /-statistic is presented for each year of five-year period and “all years” The various /-statistics are shown under the ‘all years’ column Companies are classified asSLCs or GLCs if SCIC or Government Ownership is equal to or greater than 20% of total shares GLCs are companies owned by state excluding SCIC-Linked SD is stand-for StandardDeviation.

Table 5.4 Benchmarking of SLCs and GLCs on Performance Ratios, Privileges and Business Risk (continue)

Panel A 2014 2015 2016 2017 All Years Variables SLCs GLCs T-Test SLCs GLCs T-Test SLCs GLCs T-Test SLCs GLCs T-Test SLCs GLCs T-Test Tobin's Q 1.539 0.936 0.604% ** 1.868 0.953 0.916*** 2.429 1.03 1.390***% 2.274 1.037 1.237*** 1.526 1.017 0.508***

- * _ xk _ KX _ Kx interes to Net 0.01 0.018 0.008 0.004 0.015 0.011 0.004 0.014 0.010 0.005 0.015 0.010 0.016 0.019 -0.003*

EBITDA SD 0 Cash Flows SD 0.033 0.052 -0.019 0.031 0.051 -0.019 0.028 0.056 -0.028 0.029 0.058 -0.029 0.032 0.046 -0.014***

Asterisks (***, ** and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively This table compares the means of various performance measures of SOH-Linked Companies to GLCs The mean difference /-statistic is presented for each year of five-year period and “all years” The various /-statistics are shown under the ‘all years’ column Companies are classified asSLCs or GLCs if SCIC or Government Ownership is equal to or greater than 20% of total shares GLCs are companies owned by state excluding SCIC-Linked SD is stand-for StandardDeviation.

DuPont model is used to conduct detailed analysis for ROE including three components profit margin (net income on sales — ROS), asset turnover (sales-to-total assets) and equity multiplier (total assets-to-equity) The main drivers for ROE of SLCs are equity multiplier in 2 comparison groups and asset turnover more for SLCs vs non-GLCs which significant at the 1% level Since asset turnover demonstrates how good a company is using its assets, higher values appear to imply that SLCs make good use of their assets as non-GLCs This result is even better than Ang and Ding (2006) study on Temasek The equity multiplier revealed that SLCs have lower asset- to-equity ratios than both GLCs and non-GLCs This is compatible with lower leverages (total debt to total assets and long term debt to equity) which are significant at 1% level at both 2 comparison groups A close examination revealed that SLCs have higher asset levels caused by SLCs maintain larger excess cash SLCs preserve a significantly higher cash-to-assets ratio than GLCs and non-GLCs at 1% and 5% levels Ready cash allows SLCs to fulfill greater interest payments and unexpected cash shortfalls The sampled data is collected since 2009 which is one year after 2008 global financial crisis This could be explained that SLCs maintain their own cash reserves to combat the crisis and try not be dependent on other sources after crisis regarding SLCs could adopted higher management standards and that SLCs are expected to be profitable and self-directed by its main objectives as stated in its article of association as “to make profit, preserve and develop state owned capital invested in SCIC and SCIC's capital invested in other enterprises, and fulfill other tasks assigned by the State” and “to renovate the management methodology and improve the management efficiency of state capital invested in companies” which mentioned in Decision 151/2005/QD-TTg on 20/06/2005 by the Prime Minister.

SLCs have higher Sales to Total Assets ratio which is significant at 1% in comparison with non-GLCs, as a result, caused higher expense to assets and expense to sales ratios at SLCs in comparisons to GLCs and non-GLCs The higher cash holdings (higher cash to total assets and higher cash to net assets are significant at 5% level) lead to higher capital expenditures investment ratios (capital expenditure to total assets and capital expenditure to sales) These ratios indicate that SLCs spent more on investment than other counterparts.

In addition, Table 5.6 provides the same comparisons for GLCs and Non-GLCs In general, GLCs are not good in market base performance in which Tobin and Market- To-Book ratio is lower significant at 1% level GLCs are also have lower expense to tota assets and expense to sales which are significant at 5% level This demonstrates GLCs are receiving subsidiary from Government regarding the leverage is higher.

GLCs have less cash in which cash to total assets and cash to sales are significant at 1% level GLCs’ EBITDA Standard Deviation, recorded for operations risk, which is significant lower for GLCs.

Liquidity of financial companies are different from non-financial firms regarding these companies are related to high leverage Their high leverage, however, does not have same meanings as of high leverage of non-financial firms (Dittmar and Mahrt-Smith, 2007) Therefore, excluding financial companies would provide more accurate comparisons, especially on the leverage criterion.

Table 5.7 presents t-test comparisons including SLCs versus GLCs and non-GLCs excluding financial companies These apply same criteria as in Table 5.4 and 5.5.

Sample is divided into 3 groups basing on ownership of major owners with 20% control threshold meaning SLCs are companies owned by SCIC at or above 20% The same applied for GLCs and the last group includes companies which is neither SLCs nor GLCs GLCs are companies owned by state excluding SCIC-Linked The results are consistent with previous findings presented in Table 5.4 and 5.5.

5.5 Regression Results 5.5.1 Impacts of Ownership Structure, Board Characteristics on Firm Performance

Previous studies show that investors appreciate companies with good corporate governance (Felton et al., 1996) This section would determine if SOH helps increase firm value by exploring relationship between corporate governance measurement and firm performance Regression model to determine the nature of the relation between ownership structure and firm performance includes:

Performance; = Po + Vie Ownership Structure; ae B.Board Characteristics; +

Ye, B Control Variables, + fị HOSH Dummy + Yiku12 Bndustry, +

Regression model (1) is demonstrated in details by model (1A) as below:

{TOBIN |MB}y, = Bo + BiSCIC Ownership, + f2Government Ownership; + ;Family

Ownership; + B,Foreign Ownership; + BsBoard Independence; + Bsluality;, + BGrowth Rate, + 6sLeverage, + foSize; +B Board Size, + Bị HOSH Dummy

The ordinary least squares regression technique (OLS), Fixed Effects (FE), Random Effects (RE), Feasible Generalized Least Square (FGLS) and Panel-Corrected Standard Error (PCSE) are used to test the hypotheses Each Model is regressed with 3 estimation methods (OLS, RE and FE) Details of regressions could be referred to Appendix 4 However, the tests for heteroscedasticity indicate that there is evidence of heteroscedasticity in model 1A as described in Table 5.8 below.

Table 5.4 Tests for Heteroscedasticity of Model 1A Estimation Test Chi2 Prob > chi2

OLS Breusch-Pagan / Cook-Weisberg 305.39 0.0000***

RE Breusch and Pagan Lagrangian 861.45 0.0000***

Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively

Furthermore, the Wooldridge test is performed to test for autocorrelation in panel-data model The test indicates that there is evidence of autocorrelation of study’s data.

Table 5.5 Test for Autocorrelation for Model 1A Test Chi2 Prob > chi2 Wooldridge 626.166 0.0000***

Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively

Tests of heteroscedasticity and autocorrelation show that the regressions of POLS, RE and FE delivered unreliable estimations Thus, FGLS & PCSE regression is used to fix the problems.

Table 5.6 Impacts of Ownership Structure and Board Characteristics on Firm Performance

Robustness ChecK 110111 111 1 11 n1 1 ng ng re 125

Robustness check is performed to examine whether regression coefficients are plausible and robust by adding or removing repressors If signs and magnitudes of the coefficients are plausible, estimated regression coefficients are reliable (Lu and White, 2014).

First, robustness check would be performed in sub-periods to ensure the reliability of results over time The sample would be divided into 2 sub-periods including 2009-

2012 as the period of global financial crisis and 2013-2017 as the period of recession which could have impacts on firm performance The results indicate that relationship between SOH ownership and firm performance does not change over the time in comparison to results in main regressions The below table describes the regression results of Model 1 while results of others could be found in Appendix.

Table 5.12 Robustness Check by Sub-periods Regressions of Ownership Structure and Board Characteristics

Model 1A Regressions’ results of Tobins’Q (7obin) and Market to Book (MB) by FGLS Independent variables of study include S Own, G Own, F Own, For Own, B Indep and D Dual representing for SCIC ownership, Government Ownership, Family ownership, Foreign ownership, Board Indepdence and Duality Control variables includes Growth rate (Growth), Leverage (Lev), Firm Size (Size), Board Size (B_ Size), Stock Exchange listed (D_ Hose), Industry is classified by ICB and Year (Year) Sample includes 242 firms for the period of 2009 to 2017 forming 2,137 observations divided into 2 sub-periods.

(1) (2) (3) (4) tobin mb tobin mb Ss own 0.450** 0.816* 1.135*** 0.973**

(0.0498) (0.0733) (0.0500) (0.0778) Industry Yes Yes Yes Yes Year Yes Yes Yes Yes Constant 0.989 0.337 -0.366 0.953

(0.651) (1.012) (0.898) (0.980) Observations 966 966 1,171 1,171 Number of id 242 242 242 242 Standard errors in parentheses Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels

Second, follow up with Krafft et al (2013) and Gompers et al (2003) to validate the results of regression models, Industry-Adjusted Performance Variables are created to examine for the relationship between firm performance and ownership structure and board characteristics The Industry-Adjusted-Tobin’s Q is designed by Tobin’s Q value minus the median Q for each industry The results indicate that relationship between firm value and ownership structure does not change with Industry-Adjusted- Tobin’s Q in comparison to Tobin’s Q results in main regressions Same methodology is applied for MB The results are described in below table.

Table 5.13 Robustness Check by Industry-Adjusted-Performance Regressions of Ownership Structure and Board Characteristics

Model 1A Regressions’ results of Tobins’Q (7obin) and Market to Book (MB) by FGLS Independent variables of study include § Own, G Own, F Own, Fr Own, B Indep and D Dual representing for SCIC ownership, Government Ownership, Family ownership, Foreign ownership, Board Indepdence and Duality Control variables includes Growth rate (Growth), Leverage (Lev), Firm Size (Size), Board Size (B Size), Stock Exchange listed (D_ Hose), Industry is classified by ICB and Year (Year) Sample includes 242 firms for the period of 2009 to 2017 forming 2,137 observations.

(0.0382) (0.0736) Industry Yes Yes Year Yes Yes Constant -1.011 -0.353

Standard errors in parentheses Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively

Besides, we have run similar regressions with the group of companies with State's capital, including SLCs and GLCs This sample does not include firms without State's capital, which is non-GLCs firms The result shows that the impact of SCIC's ownership on firm market performance is positively significant The coefficient of

SCIC's ownership is also larger than that of State's ownership in GLCs, which mean that SCIC's ownership marginally adds values to GLC's firms Details are presented in Appendix 6. Š.7 Endogeneity Check

Given endogeneity was argued by several researchers such as Demsetz (1983) and Demsetz (1983) although this view was being challeged by Gugler & Weigan (2003), Instrumental variable (IV) method was performed to control for potential endogeneity similar to approach of Hermalin & Weisbach (1991), Gugler & Weigan (2003) and DzUanié (2012) in which lagged explanatory variable is used as instrument for ownership structrure (Hermalin & Weisbach, 1991; Gugler & Weigan; 2003;

DzL 'anié, 2012) Consistently, IV with lagged ownership instrumental approach shows that S_Own still affects performance with details as below.

Table 5.14 Impacts of Ownership Structure and Board Characteristics on Firm Performance by IV (2SLS)

Model 1A Regressions’ results of Tobins’Q (Tobin) by 2SLS (IV) Ownership variables of study include S Own, G Own, F Own, Fr Own, B Indep and D_ Dual representing for SCIC ownership, Government Ownership, Family ownership, Foreign ownership, Board Indepdence and Duality Control variables includes Growth rate (Growth), Leverage (Lev), Firm Size (Size), Board Size (B_Size), Stock Exchange listed (D_Hose).

IV VARIABLES tobin Ss own 0.921***

(0.0340) Industry Yes Year Yes Constant 0.371

Standard errors in parentheses Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively

Furthermore, lagged values of ownership variables are added into regression model following approach of Xie & Chen (2016) and the result indicates that lagged values have minor influence on firm performance and the outcomes are consistent with PCSE results Meanwhile, there is little simultaneously between ownership structure and firm performance The detail is described in below table.

Table 5.15 Impacts of ownership structure and board characteristics on Firm Performance by Lagged Ownership

Model 1A Regressions’ results of Tobins’Q (7obin) by lagged ownership variables Ownership variables of study include S Own, G_Own, F Own, Fr_Own, BIndep and D_ Dual representing for SCIC ownership, Government Ownership, Family ownership, Foreign ownership, Board Indepdence and Duality Control variables includes Growth rate (Growth), Leverage (Lev), Firm Size (Size), Board Size (B_ Size), Stock Exchange listed (D_ Hose).

Asterisks (***, **, and *) denote statistical significance at the 1%, 5%, and 10% levels, respectively

PCSE VARIABLES tobin Lagged.s own 1.036***

(0.0526) Industry Yes Year Yes Constant 1.420*

(0.826) Observations 1,899 Number of id 242 R-squared 0.165

5.8 Summary of ChapterChapter 5 presented the analysis results of empirical data collected on companies listed on the HOSE and HNX.

In comparison analysis, SLCs deliver superior returns and are valued more highly than GLCs and non-GLCs SLCs perform better than GLCs and non-GLCs in many performance measures and do not seem worse in other measures Respectively, they are more highly valued.

In regression analysis, firm performance is measured by Tobin’s Q and Market-to- Book ratio The regressions are estimated by OLS, RE and FE However, tests of heteroscedasticity and autocorrelation expose the regressions of POLS, RE and FE delivered unreliable estimations Thus, FGLS and PCSE regressions are used to fix the problems.

Regressions on Model 1 indicate that ownership structure, board characteristics and firm performance have relationships in which SCIC Ownership, Foreign Ownership, Board Independence, Size and Exchange are found to have positive significant impacts on firm performance while Leverage shows a negative effect Government Ownership, Family Ownership, Size, Duality and Board Size do not impact on firm performance in this Model.

Furthermore, [IV method was performed to exmamine for potential endogeneity in reasearch model whereby lagged variables are defined as instrumental variables in estimator to explore one-way relationship within the model The results are consistent with PCSE outcomes in which S Own and Fr Own are found to have positive relationship with firm performance This result implicits that SCIC has a one-way causal relationship with firm performance Besides, lagged ownership values are included into regression model and the result indicates that lagged variables have minor impact on results meaning there is little simulteneously between ownership structure and firm performance.

Summary of Chapter cccccccccccccsssssssssssceeeeecceeeeeseessssssaeeeeeeeeeeeeesseeseegas 129

In comparison analysis, SLCs deliver superior returns and are valued more highly than GLCs and non-GLCs SLCs perform better than GLCs and non-GLCs in many performance measures and do not seem worse in other measures Respectively, they are more highly valued.

In regression analysis, firm performance is measured by Tobin’s Q and Market-to- Book ratio The regressions are estimated by OLS, RE and FE However, tests of heteroscedasticity and autocorrelation expose the regressions of POLS, RE and FE delivered unreliable estimations Thus, FGLS and PCSE regressions are used to fix the problems.

Regressions on Model 1 indicate that ownership structure, board characteristics and firm performance have relationships in which SCIC Ownership, Foreign Ownership, Board Independence, Size and Exchange are found to have positive significant impacts on firm performance while Leverage shows a negative effect Government Ownership, Family Ownership, Size, Duality and Board Size do not impact on firm performance in this Model.

Furthermore, [IV method was performed to exmamine for potential endogeneity in reasearch model whereby lagged variables are defined as instrumental variables in estimator to explore one-way relationship within the model The results are consistent with PCSE outcomes in which S Own and Fr Own are found to have positive relationship with firm performance This result implicits that SCIC has a one-way causal relationship with firm performance Besides, lagged ownership values are included into regression model and the result indicates that lagged variables have minor impact on results meaning there is little simulteneously between ownership structure and firm performance.

The interaction analysis in Model 2 shows that the more dominant the SCIC, the greater its impact on a company's performance improvement Moreover, Model 2 also found that if the dominant owner is GLC, company performance would be negatively impact.

In general, the results support 2 hypotheses HI, H3 and H4 while do not support H2 and HS The controlled variable firm leverage and listed on HOSE have positive impacts on firm performance while others are found do not have significant impacts.

Industry and year controlled variables are found to have impacts on firm performance indicating that firm performance could be affected by external environmental factors.

Regression on ownership structure and corporate cash holdings shows that SCIC ownership and foreign ownership have positive significant relationships with cash holdings Government ownership and Family ownership is found do not have negative relationship with cash holdings This finding supports evidence from Opler at al.

(1999) in which firms do well tend to hold more cash than predicted regarding cash holdings allows firm pursuing investments opportunities and reduces the risk of financial distress according to trade-off model (Ferreira and Vilela, 2004; Martinez- Sola et al., 2018) The results support Heyl.

SCIC ownership and foreign ownership, moreover, significantly increase the value of cash holdings: the coefficient on the interaction variable between excess cash and these types of ownership are consistently positive and significant for both level of excess and change of excess cash models The results indicate that the value of excess cash is statistically and economically significantly greater if the firm is managed by SCIC or foreign investors It demonstrates that good performance companies not only hold more cash but also this excess cash has greater value The results support for

There is argument that quality of corporate governance improves firm value Dittmar and Mahrt-Smith (2007) provided empirical evidence that good corporate governance would increase value of cash reserves, or in other words, good corporate governance would improve the value of the company by improving the use of cash holdings(Dittmar and Mahrt-Smith, 2007; Seifert & Gonenc, 2017) This study extended the findings of Dittmar and Mahrt-Smith (2007) in which SLCs are found to hold more cash and the shareholders appreciate the cash hold by the SLCs Without evidence thatSLCs have special privileges and lower business risks, the better shareholder value of cash in SLCs is revealed to come from better corporate governance and his increase the firm value by improving the use of cash holdings.

DISCUSSIONS eee cecccsscceesssseeceseseeecesesseeceseeseeeeeseeesessesaeeees 133 “acc on ha 4 a

Summary of Main Findings from SampÌes . - 55555 << s+++++++ssssss 133 0)5720490).70Ì0.040:i0.00090) 55

This study conducts an empirical study to analyze of impacts of ownership structure on firm performance of listed companies in Vietnam Standing out of the recent studies for Vietnam, this thesis not only retests the underlying assumptions of these studies but also includes two completely different aspects First, this thesis analyses in detail the important role of SCIC, the State-owned Holding Company Secondly, it updates the data used in studies The thesis contributes to the limited empirical understanding of how SOH operates as a shareholder in SOH-Linked Companies (SLCs) Although Vietnam is different from Singapore and Malaysia, SOH, however, are found to have positive impacts on firm performance which ascertains the role of SOH is entirely homogeneous regardless corporate governance environment.

An important objective of this study is to compare various financial and market performance of SCIC linked companies (SLCs) with other GLCs and non-GLCs,which have different ownership structure and the key difference being government ownership The samples show that, on average, SLCs deliver superior returns and are valued more highly than GLCs and non-GLCs (Tobin’s Q, market-to-book, ROE andROA) DuPont analysis indicates that SLCs have lower leverage and maintain significantly higher cash-to-assets ratio than GLCs and non-GLCs Rich cash holding allows SLCs to fulfill greater interest payments and unexpected cash shortfalls SLCs perform better than GLCs and non-GLCs in many performance measures, and more importantly, they have higher market values The results suggests that investors inVietnamese market value the corporate governance standards of SLCs than otherGLCs or non-GLCs Given in GLCs, the government plays a dual role of both shareholder and regulatory of the corporate sector, and as a result, GLCs must be responsible for non-commercial objectives besides the profit-maximization responsibilitie leading to conflict of interests agency problem The research results based on the samples show that SLCs’ performance is more effective in managing its invested companies than GLCs.

In more details, in this study, to answer for question on the role of SOH model in Vietnamese market, the ownership of SCIC is studied on 242 listed firm This study found that SOH Ownership and Foreign Ownership have positive impacts on firm performance The interaction analysis revealed that the more dominant the SCIC Ownership, the greater its impact on a company's performance improvement.

Moreover, interaction analysis also found that if the dominant owner is GLC, company performance would be negatively impacted These findings contributes understanding on the role of SOH into current literature, especially on agency theory perspective.

SOH, above all its mandates, is a strategic investor with profit maximization orientation (Kuznetsov & Murav’ev, 2001; Sam, 2010; Kim & Chung, 2018) Acting as a insitutinal investor, agency problem due to conflict of interests could be overcome given SOH has strong incentives to implement better governance mechanisms (Sam, 2013) SOH, besides, servers as a buffer between government and SLCs (Sam, 2007;

Chen, 2016) and this structure alleviate agency problems in state capital firms including conflicting objectives, political interference and lack of transparency (Wong, 2004; Peng et al., 2016; Nurgozhayeva, 2017) The findings obtained from samples are reference to policy makers, investors and relevant stakeholders to figure an effective corporate governance framework in organizations.

Having been a centrally controlled economy, it is revealed a dominant role of the State in Vietnamese economy The state capital firms remains significant despite a steady decline in their contribution to GDP growth (Taussig et al., 2015) State capital firms are found to pursue multiple — and conflicting — objectives in literature (Wong, 2004;

Lin, 2012; Chen, 2013; Peng et al., 2016; Nurgozhayeva, 2017) In general, state capital firms do not only have commercial goals but they have also to serve social objectives such as providing jobs and serving public services (Kamal, 2010; Peng et al., 2016) This leads to contradictory objectives in state capital firms (Wong, 2004;

Chen, 2013; Peng et al., 2016) Therefore, there is a concern of state capital firms’ performance which many studies have found that state ownership does not only produce superior firm performance but also often linked to low efficiency (Hu et al., 2009) This thesis, however, puts into question such arguments taken into consideration the role of SOH.

Similar to the results found in Singapore with Temasek model where better governance exists (Ang and Ding, 2006), this study suggests that SOH is an potential model to mitigate the problems of state capital firms governance The regressions found that SCIC ownership has a positive relationship with firm performance SOH is a strategic investor with profit maximization orientation (Kuznetsov & Murav’ev, 2001; Sam, 2010; Kim & Chung, 2018) and the holding structure seems to well serve the purpose of resolving the first two problems at state capital firms as the holding structure would be able a middle layer shielding invested companies from politics and government intervention while transparency can be best improved by opening access of ownership to the public (Wicaksono, 2008) Placing state capital firms under the control of an SOH rather than the direct ownership of the state might reduce the conflicts when the state serves both roles of shareholder and regulator (Chen, 2013).

SOH acts as a safety valve between a regulator and a regulated firm (Hamdani and Kamar, 2012) This would allow the government the flexibility to deal with a particular firms or industry, and may help avoid a dilemma when regulatory enforcement troubles the government’s interests as a shareholder (Chen, 2013).

Following model of Chen at al (2006), the interaction analysis, moreover, revealed that the more dominant SCIC Ownership, the greater its impact on company's performance.

This study, however, does not found significant correlation between non-SOH state ownership with firm performance This result is inconsistent with results in literature which shows that state ownership is linked with low efficiency and ineffective performance (Bai et al., 2004; Ding et al., 2007; Nee et al.; 2007; Phung and Hoang,2013; Tran et al., 2014) Tran et al (2014) used different types of companies ranging from private to public companies and found that state capital firms are facing with three main challenges The study’s result is partially compatible with study of Phung and Hoang (2013) as they found that state ownership could improve firm performance when the ownership is not concentrated and vice versa This could be explained as a result of variety of special privileges granted to state capital firms and give them a support against non-state competitors (Taussig et al., 2015) State capital firms, moreover, are receiving beneficial policies and favor from the government which give them competitive advantages over private entities (Taussig et al, 2015) The improvement of corporate governance in recent years, the anti-corruption campaign from government and the privatization acceleration could be other explanations to this finding as well.

Recently, conflicts between Controlling Shareholders and Minority Shareholders causing principal—principal conflicts are taken into consideration in Asian countries where the ownership concentration is dominant (Gửnencer, 2008; Claessens and Fan, 2002; Claessens et al., 2000; Young et al., 2008; Driffield and Pal, 2007; Nam et al., 1999) Majority control gives the larger shareholders considerable power and discretion over key decisions (Stiglbauer, 2011) The efficacy of ownership concentration is a controversy of monitoring versus expropriation role Concentration ownership is believed to limit agency problem as higher concentration gives large shareholders stronger incentives and greater power at lower cost to monitor management (Hu and Izumida, 2008) but interests of large shareholders could be diverged from minority shareholders’ benefits (Hu and Izumida, 2008) and controlling shareholders could exploit the interests of minority shareholders (Hu et al., 2009).

Methodologies to measure ownership concentration of almost studies after the research of Demsetz and Lehn (1985) accumulate the ownership five, ten, or twenty largest shareholders However, Earle et al (2005) argued that group accumulation could conceal the interactions among large shareholders and the pattern of concentration.

The approach of measuring ownership concentration by largest blockholder is supposed to be better than group measurement Therefore, dominant ownership is used to measure ownership concentration in this study by subtracting largest ownership for opponent ownership The study found that family ownership has negative relationship with firm performance This result is contradicted to result of Anderson and Reeb(2003) in US market However, it is consistent with findings of Connelly at al (2008) and Giovannini (2010) The finding, therefore, supports arguments of Yeh et al (2001) and (Bloom and Van Reenen, 2006) in which family representation on the board leads to centralization in authority and decision-making power and as a result could increase the expropriation of non-family minority shareholders It also supports for argument of Claessens et al (2000) who argue that in family companies, unqualified members could be appointed to key positions without competition Moreover, because of close relations and informal linkages, family managers are less to be monitored (Young et al, 2008).

Implications for Literature based on findings from study’s samples

The findings, based on 242 companies samples, found a positive relationship between SLCs and firm performance which provides a different understanding of state capital firms The empirical results differ significantly from adverse argumentation of state ownership in most existing literature SLCs are not linked with low performance as found in previous studies on state capital firms (Bai et al., 2006; Hu et al., 2009; Tan et al., 2015) This can be explained by the fact that SOH is considered as a mechanism to resolve conflict of interests in agency theory SOH is an intermediary agent that acts like a direct investment holding arm of countries’ governments with profit maximization orientation (Sam, 2008; Kim & Chung, 2018) The study finding addressed the first question of drivening by the foremost profitability objective, is the performance of SLCs better than GLCs? Given SOH is a buffer to reduce agency problem (Kim & Chung, 2018), this study contributes an empirical result to consolidate the role of SOH in existing literature.

SOH, first of all, would act a role similar to an institutional investor and have stronger incentives to maximize firms’ value As a result, agency problem due to conflict of interests could be overcome (Sam, 2013) In addition, acting as an institutional investor with large enough ownership, SOH can use their rights to place pressures on managers in improving corporate governance, transparency, which lead to lower agency problem (Sam, 2013) This result contributes to an experimental outcome to agency theory on institutional perspective SOH model is supposed to overcome agency problems existing in state capital firms and serve as a one-stop agency to resolve problems related to its SLCs (Kim & Chung, 2018).

Secondly, SOH is assumed as an effective model to mitigate agency problem exist between government and the firm The holding structure seems to well serve the purpose of resolving the first two problems of objectives confliction and political interference at state capital firms given the holding structure is a layer shielding the SOH-invested companies from politics and government intervention while transparency can be best improved by opening access of ownership to the public (Wicaksono, 2008; Sam, 2013) The positive correlation between SOH and firm performance suggests the effectiveness of this model even in a lack of good corporate governance environment This model would be potentially compatible with Wong’s (2004) and Peng at el (2016) three pillars of state capital firms reform: avoiding conflicting objectives, minimizing political intervention and improving transparency (Wong, 2004; Peng et al., 2016) Agency theory perception of state capital firms is pessimistic since conflicting interests arising in monitoring and controlling by government institutions (Wicaksono, 2008), the finding of this study on SOH role questions literature which is adverse to state ownership and contributes the result from Vietnamese context to the limited empirical understanding of how SOH operates as a shareholder in SOH-Linked Companies (SLCs).

Besides, SCIC ownership and foreign ownership, specifically, is found to increase the value of cash holdings: the coefficient on the interaction variable between excess cash and these types of ownership are consistently positive and significant with firm value based on existing samples This addressed the second research question of does SOH help increase firm value through improving value of cash holdings in SLCs? The findings express that SOH-invested companies not only hold more cash but also this excess cash has greater value This outcome supposes a relationship between SOH and good corporate governance derived from findings of Dittmar and Mahrt-Smith (2007) and Seifert & Gonenc (2018) who found poor corporate governance could waste cash and destroy firm value while good corporate governance could utilize cash holding to have better performance (Dittmar et al., 2003) Comparative analysis shows that SLCs do not receive special privileges, the better value of cash in SLCs is assumed come from a better corporate governance Given agency theory suggests that poor corporate governance can weaken managers’ fiduciary responsibilities and allow self-interested managers to entrench and engage in power concentration (Jensen, 1986), firms with better governance would reduce this managerial agency problem and provide shareholder more power in disciplining managers (Hu & Izumida, 2008) SLCs performance, therefore, suggests that SOH ownership has effective role and better governance This finding contributes empirical understanding on role of SOH in agency theory under cash holdings perspective This finding also contributes other empirical outcome to the literature of corporate governance and provides an potential evidence for SOH as a mechanism in improving corporate governance standards and this affects firm performance.

Besides main findings on SOH, based on study’s samples, empirical findings also contribute to existing literature on different perspectives Instead of traditional principal—agent conflicts emerged in most studies in developed economies, principal— principal conflicts have been identified as a major concern of corporate governance in emerging economies, especially in Asian Principal-principal conflicts between controlling shareholders and minority shareholders result from concentrated ownership, extensive family ownership and weak legal protection of minority shareholders (Young et al., 2008) There could be expropriation of minority shareholders from large shareholders (Claessens et al., 2000) The finding on family ownership has negative impact on firm performance provide another support to the argument of family representation leads to authority and decision-making power centralization and as a result could increase the expropriation of non-family minority shareholders (Yeh et al., 2001; Bloom and Van Reenen, 2006).

This study, moreover, contributes empirical evidence to have more understanding on cash policy through relationship between ownership structure and value of cash holdings SCIC ownership and foreign ownership are found have positive significant relationships with cash holdings and provide an empirical evidence to trade-off model and precaution motive Like Opler et al (1999), Ferreira and Vilela (2004) andMartinez-Sola et al (2018) whose studies supports for trade-off model, SCIC and foreign ownerships favor the arguments that successful companies tend to hoard more cash Regarding trade-off model, cash holding reduces the risk of financial distress as it provides a safe replacement for unexpected expenses or external financial constraints Secondly, cash holding allows firm pursuing investments opportunities in times when external funds are not available Finally, cash holding reduces costs of raising external funds or liquidating current existing assets (Ferreira and Vilela, 2004).

This result is aligned with precautionary motive to hold cash of business given firms hold cash as precaution to cover unforeseen potential necessities (Opler et al., 1999;

Ferreira and Vilela, 2004; Martinez-Sola et al., 2018) Besides, it also provide another empirical result to support transaction motive assuming that companies hold cash for operating expenses to meet payment responsibilities (Opler et al., 1999; Bates et al.,2009).

Implications for Policy Makers based on findings from study’s samples

This study provides policy makers, managers, investors and stakeholders in Vietnam empirical findings on the influence of the ownership structures and board characteristics to firm performance Therefore, regulators have another reference during the policy making process on corporate governance to be suitable to Vietnam’s conditions in which SOH model could be encouraged and replicated It also helps managers make adjustments, improvements in corporate governance at company level in order to achieve better firm performances The study also helps the investors and other stakeholders understanding better the problems of corporate governance in Vietnam Therefore, it can help them in making decisions in investments, choosing board directors, or making corporate governance policy.

Firstly, the State Securities Committee should consider cooperating with international organizations such as the World Bank and IFC to build effective corporate governance framework focusing on ownership structure and board characteristics which are found to have significant impacts on firm performance As a result, the positive related factors should be encouraged like SOH or Foreign ownership while the negative related factors should be monitored such as family ownership The encouragement or monitoring could be implemented by constraints in corporate governance framework or regulations.

The role of SOH is found to have positive impact on firm performance in condition ofVietnam This is the answer for the concern of the workability of SOH model in the lack of good corporate governance environment of Vietnam Temasek is found to be effective in Singapore where good and clean governance exist and SOH model is also found to have positive effect in Vietnam context with SCIC model Based on study’s samples, SCIC is found to have successful likenesses of Singapore’s Temasek.

Therefore, it could be a reference to build an appropriate model for equitization process in Vietnam SCIC or other State-owned Holding Companies could provide a model to be considered to improve performance of state capital firms or GLCs which are believed as ineffective and full of corruption However, Temasek is operating in developed and transparent financial market and managed by professional directors which is not easy to imitate It is a challenge in building a successful SOH model and requires a lot of efforts at all levels of management in the formulation of policies and selection criteria for executive team in terms of Vietnam SCIC has relationship with

Temasek and the sharing of operating experience of the two companies'” also will be useful for improving the operation of SCIC and for the Vietnamese fund management companies in general.

There is negative correlation between family ownership as the largest owner and firm performance in Vietnam practice This is explained as centralization in authority and decision-making power, unqualified key positions appointments and less monitoring.

Therefore, it requires more attentions on regulator to build-up mechanism to monitor family company to protect minority shareholders.

Secondly, investors could support regulatory during corporate governance framework development, especially for foreign investors who brings a wide-range of knowledge from international markets An effective corporate governance framework, on the vice versa, would support investors in decision-making to choose effective company.

Moreover, investors need to be more active in monitoring invested companies to drive these companies to be more efficient and transparent.

Thirdly, SCIC and foreign investors are found to have positive relationship with cash holdings and firm value This result is supposed to derive from good corporate governance given poor corporate governance could waste cash and destroy firm value while good corporate governance could utilize cash holding to have better

'© SCIC (2016) Tăng cường hop tác chiến lược giữa SCIC va Temasek SC/C Retrieved from https://www.scic.vn performance (Dittmar et al., 2003) Therefore, an effective corporate governance framework would contribute to improve company performance and roles of SCIC and foreign investors should be taken into consideration.

Last but not least, every company needs to be aware of the important role of corporate governance and to force themselves to meet the strictest criteria to improve the business Efficient company would receive attention from investors and stakeholders and its capital needs would easily be met to support for its growth.

Applying quantitative method on panel data of listed companies in Vietnam during 2009-2017, this study found that SOH ownership have positive impacts on firm performance which provides an empirical contribution to existing literature on agency theory and corporate governance perspectives Similar to the results found in Singapore with Temasek model where better governance exists, the result of Vietnam shows that SOH is supposed to be an appropriate model to mitigate the problems of state capital firms The finding in Vietnam strengthens the findings in other countries that SOH model mitigates the agency problem SOH is assumed to reduce agency costs by serving as an institutional investor who has strong incentives to maximize SLCs’ value and increase accountability of SLCs Acting as an institutional investor with large enough ownership, SOH can use their rights to place pressures on managers in improving corporate governance, transparency, which lead to lower agency problem SOH model overcomes agency problems of state capital firms and serve as a one-stop agency to resolve problems related to SLCs.

SOH is a model in which government does not directly manage the enterprises as traditionally model An investment company is established and represents the ownership of the government in companies As an active shareholder, SOH is playing a key role in raising corporate governance standards for SLCs to mitigate agency problem and allowing for more reliable long-term engagements between SOH andSLCs.

Research results have positive implications in conditions in Vietnam SCIC is SOH representing interests in SLCs and becomes a government’s strategic investor to maximum value and sustainable returns The interaction analysis revealed that the more dominant the SCIC Ownership, the greater its impact on a company's performance improvement As Vietnam is a premature capital market economy with developing corporate governance framework, this finding strengthen the findings in other countries and assert the role of SOH in condition of week corporate governance environment.

The findings indicate SCIC ownership and foreign ownership have positive significant relationships with cash holdings This supports the argument that firms do well tend to hold more cash than predicted given cash holdings allows firm pursuing investments opportunities and reduces the risk of financial distress following trade-off model and precaution motive SCIC ownership and foreign ownership, specifically, significantly increase the value of cash holdings: the coefficient on the interaction variable between excess cash and these types of ownership are consistently positive and significant The result indicates that the value of excess cash is statistically and economically significantly greater if the firm is managed by SCIC or foreign investors The outcome assumed a good corporate governance existence given previous studies found that good corporate governance would increase firm value through improving value of cash holdings.

The research has some limitations Data was collected from public data sources such as annual reports and the third-party sources This information was assumed to be accurate and correct It would be limited in some issues related to the accuracy or reliability Moreover, data consists of 242 companies listed on the HOSE and HNX for 9 years from 2009 These companies listed before 31/12/2008 Other companies listed after this date are not included into portfolio The data therefore is not large Small data sets may affect the accuracy of the model.

This study initially identified the impact of SOH ownership structure on firm performance through regression models as well as through cash holdings model.

However, SCIC's role in improving the efficiency of state capital investment should be comprehensively explored in more details in subsequent studies.