Does owership structure affects firm performance the evidence from vietnam stock market

On the other hand, we also found an inverted U-shaped relationship between state ownership and Tobin’s Q, indicating that partial privatization could be an efficient way to improve firm

1

Does ownership structure affect firm performance? Evidence from the

Vietnamese stock market Lai Hoanga,b,* , Cuong Nguyena, Baiding Hua

a Faculty of Commerce, Lincoln University, New Zealand

b School of Banking and Finance, National Economics University, Vietnam

This version: 17 Oct 2016

to external monitoring in corporate governance On the other hand, we also found an inverted U-shaped relationship between state ownership and Tobin’s Q, indicating that partial privatization could be an efficient way to improve firm performance

Keywords: ownership structure, managerial ownership, block ownership, state ownership, firm

performance

JEL Codes: G32, G34.

* Corresponding author Tel: + 64 273593522, + 84 944656365

Email address : trunglai.hoang@lincolnuni.ac.nz

2

1 Introduction

The separation of ownership and control in the modern corporation model leads to the well-known agency problem (Fama & Jensen, 1983; Jensen & Meckling, 1976), i.e managers have incentives to exploit firm’s resources to serve their private purposes rather than that of shareholders (Jensen, 1986) Meanwhile outside owners’ incentives in supervising managers is weaken by the “free rider problem” Therefore, highly concentrated ownership structure is expected to alleviate both agency problem and

free rider problem as it aligns interest of managers and outside shareholders (convergence-of-interest hypothesis), as well as increases efficiency of monitoring mechanisms (monitoring hypothesis) Two

common measures of ownership concentration are fraction of shares owned by the largest shareholders (block ownership) and by firm managers (managerial ownership1) Since large shareholders could affect firm strategies and operations through their significant voting rights and controlling power, block ownership represents the ability and motivation of shareholders in monitoring managers (external pressure) On the other hand, managerial ownership reflects inner incentives of the management team itself in operating firm effectively (internal motivation) High ownership concentration enhances both external pressure and internal motivation, therefore could positively affect firm performance

However, Shleifer and Vishny (1997) argued that the agency problem could also exist among shareholders Because of their significant influence, large shareholders could abuse that power to

exploit company’s resources and harm the efficiency of the firm as a whole (expropriation hypothesis)

The expropriation could be in various forms, such as pecuniary or unfair related party transactions, i.e selling firms’ products to other companies owned by the controlling shareholder at lower-than-market price, or buying at higher-than-market price (Barclay & Holderness, 1989) On the other hand, Shleifer and Vishny (1989) stated that if managers own sufficient shares to classify themselves as significant large shareholders, they may have enough power and influence to ignore both shareholders’ and market monitoring mechanisms to entrench their employment and salary, as well as stay on the job even if they are no longer applicable As monitoring mechanisms become ineffective, agency problem become severe Thus, it is possible that the increase of managerial ownership, especially at high level,

negatively affects firm performance (entrenchment hypothesis)

1The term managerial ownership is normally used interchangeable with insider ownership or board ownership

3

The relationship between state ownership and firm performance has also attracted significant concern The last few decades witnessed a tidal wave of privatization in transition economies of the former Soviet Union and Western Europe Although the efficiency of these privatization programs is inconclusive, it is widely accepted that from financial perspective, state-owned enterprises (SOEs) normally perform worse than private-owned counterparts The failure could be attributable to the dual principal-agent problem in SOEs: managers are agents of the state in daily decision making process, but the state is

in turn the agent of “true” owners: the voting population (Yarrow, King, Mairesse, & Melitz, 1986) As these owners are extremely diffuse, the free rider problem become even more severe than that of joint stock companies, which potentially lead to the inefficiency of SOEs However, SOEs have the advantage of receiving “helping hand” from the governments (Shleifer & Vishny, 2002), possibly in forms

of financial resources, business networks or economic contracts with government agencies As a result, the decrease of state ownership could negatively affect firm performance

This study contributes to the literature in three ways First, while most theories and empirical evidence

on the effect of ownership structure (particularly ownership concentration) on firm performance are based on the well-developed economies of the U.S and Europe, there is lack of research focusing on emerging countries To the best of our knowledge, this is the first paper considering the impact of managerial ownership and block ownership on firm performance in the Vietnamese stock market Second, the relationship between state ownership on firm performance in the context of Vietnam, which was found to be negative by Tran, Nonneman, and Jorissen (2014) and inverted U-shaped by Phung and Hoang (2013), is reexamined using an updated dataset and a more advanced method Third, instead of traditionals use of OLS, fixed effects or instrumental variables, this study employs the well-developed system GMM estimator to address the endogeneity of ownership structure, as well as the dynamic nature of the relationship between ownership structure and firm performance

2 Literature Review

Regarding to the impact of ownership structure on firm performance, not only are there conflicts in theoretical perspectives but empirical results are also inconclusive In the U.S market, most current

studies found the insignificant relationship between block ownership and firm performance, but the conclusion in case of managerial ownership is mixed The early study could be traced back to Demsetz and Lehn (1985) On the sample of 511 firms in the U.S from 1976 to 1980, they did not find a significant relationship between block ownership and accounting profit rate McConnell and Servaes (1990) on one hand confirmed Demsetz and Lehn’s result, on the other hand found an inverted U-shaped relationship between managerial ownership and Tobin’s Q using piecewise OLS Applying 2SLS to control for endogeneity, Demsetz and Villalonga (2001) reported neither impact of managerial ownership nor block ownership on Tobin’s Q The insignificant relationship between managerial ownership and firm performance was also documented in Cho (1998) and Loderer and Martin (1997)

In contrast, Chung and Pruitt (1996) and Palia and Lichtenberg (1999) reported the positive impact of managerial ownership on Tobin’s Q and Total Factor Productivity using 2SLS and piecewise OLS respectively Meanwhile, non-monotonic relationship between managerial ownership and Tobin’s Q is also found in Morck, Shleifer, and Vishny (1988), Holderness, Kroszner, and Sheehan (1999) (cubic), Himmelberg, Hubbard, and Palia (1999) (inverted U-shaped) or Hermalin and Weisbach (1991) (inverted W-shaped)

The empirical evidence on the relationship between ownership concentration and firm performance is even more inconclusive in countries outside the U.S The insignificant relationship was reported in Welch (2003) in Australia and Sheu and Yang (2005) in Taiwan Shah and Hussain (2012) also confirmed no impact of block ownership on Tobin’s Q, but the negative impact of managerial ownership was found On the contrary, positive impact of managerial ownership on firm performance was documented in Kapopoulos and Lazaretou (2007) in Greek, Li, Moshirian, Nguyen, and Tan (2007) and Liu, Uchida, and Yang (2012) in China, Fauzi and Locke (2012) in New Zealand Using IV-GMM and panel data analysis respectively, De Miguel, Pindado, and De la Torre (2004) and Short and Keasey (1999) discovered the cubic relationship between managerial ownership and firm performance in Spain and the UK However, the turning points are significantly different, i.e 35% and 70% in case of Spain, and 15% and 42% in the UK

In contrast to the U.S., the relationship between block ownership and firm performance was found to

be significant in many other countries but its signs were mixed The inverted U-shaped relationship was

reported in numbers of European countries (Balsmeier & Czarnitzki, 2015; De Miguel et al., 2004; Thomsen & Pedersen, 2000), meanwhile the U-shaped relationship was found in China by Liu et al (2012) on the sample of 970 listed firms during crisis period of 2007-2008 However, using a smaller sample of 149 Chinese listed manufacturing firms in 1999-2002, Jiang, Yue, and Zhao (2009) discovered the negative relationship, which is confirmed by Balsmeier and Czarnitzki (2015) in New Zealand On the other hand, the positive impact of block ownership on firm performance was also documented by Gedajlovic and Shapiro (2002) in Japan and Kapopoulos and Lazaretou (2007) in Greek

Regarding to state ownership, there is mounting evidence of its impact on firm performance Villalonga (2000) provided a metareview on existing empirical research comparing performance of firms with various level of state ownership Among 153 studies reviewed, 104 support the higher efficiency of private ownership, 14 against and 35 neutral In addition, there are some studies finding a non-monotonic relationship For example, in the review of Yu (2013) of empirical studies in China, 6 out of

14 studies reported the shaped relationship, meanwhile other 4 negative, 1 positive, 1 inverted shaped and 2 neutral In the context of Vietnam, Loc, Lanjouw, and Lensink (2006) compared the performance of SOEs in pre- and post-privatization period, using difference in difference (DID) method

U-to avoid selection bias They concluded that the observed improvement in profitability, sale revenue, efficiency, and employee income is truly resulted from privatization The result implies the negative impact of state ownership on firm performance Similarly, in a cross sectional comparison of more than 2,000 firms during 2004-2012, Tran et al (2014) discovered the negative impact of state ownership on firms’ profitability (ROA) and labor productivity (value added per employee) Meanwhile, Phung and Hoang (2013) found that the state ownership has an inverted U-shaped impact on Tobin’s Q and ROA

in a sample of listed companies on the HOSE and HNX (Hanoi Stock Exchange) from 2007 to 2012

Generally, listed companies in Vietnam follow two-tier corporate governance system, although there are some modifications making it slightly different with the typical two-tier structure of Germany The basic structure is visualized in Figure 1 General Meeting of Shareholders (GMS) has the power of making decision on the most important issues of the company such as long-term strategies, stock issue

and dividend According to the Vietnamese Enterprise Law 2014, GMS must be held at least once per year, within four months after the end of financial year Board of Directors (BOD) acts as an agent of GMS and has full authority to make decisions in the name of the company, especially strategical decisions that strongly affect the interest of shareholders The BOD has the right to appoint Executive Board, which is in charge of daily operations Supervisory Board is elected by the GMS and independent from both BOD and Executive Board, with the main duty of supervising these Boards in managing and operating the company While members of Executive Board could be selected from BOD, as the regulation, members of Supervisory Board must not hold any position in both BOD and Executive Board,

or be relatives of any member in those Boards

Figure 1 Corporate Governance Structure in Vietnam 2

Appointment and Removal; Supervision

Theoretically, the independence of BOD, Supervisory Board and Executive Board allow them to cooperate with, as well as supervise each other in managing and operating the company However, from the principal-agent relationship perspective, all of them share the role of shareholders’ representatives who act on behalf of shareholders to operate the firm Thus, their incentive and interest

is likely to be similar in terms of exploiting outside shareholders Therefore, in the examination of the relationship between managerial ownership and firm performance, we consider all three Boards as one unified management team, thus managerial ownership is computed as the total shares owned by all members in these Boards

2 As a unique two-tier corporate governance system in Vietnam, Board names are translated differently in different documents, which sometimes conflicts and make confused This study uses the translation system of Nguyen, Locke & Reddy (2015) to ensure consistency

General Meeting of Shareholders (GMS)

Board of Directors (BOD)

Supervisory Board

Chairperson

of BOD

Executive Board

3 Research Design

Our sample includes manufacturing companies listed on the HOSE during 2007-2015 Data is collected from companies’ annual reports, corporate governance reports and financial statements However, as ownership structure release is optional, data is unavailable in some companies, especially

hand-at the beginning of the research period On the other hand, according to Vietnamese nhand-ational listing rules, all top managers (including BOD, Supervisory Board, Executive Board and the Chief Accountant) must commit to hold 100% of their shares at least 6 months after listing day, and 50% during subsequent

6 months Therefore, firms listed in year t are included into the sample from year t+2 in order to allow sufficient time for adjustments of managerial ownership and its potential effects on firm performance to take place Based on this restriction and the availability of information, an unbalanced data of 76 out of

95 manufacturing listed companies in HOSE has been constructed, with 406 firm-year observations

Beta coefficient (β) is estimated by regressing weekly excess stock returns on weekly excess VN-Index returns of two recent years Risk free rate is weekly interbank offered rate obtained from Datastream Returns are restricted within two years allows to incorporate only recent information of firm risk, which could possibly be disturbed by out-of-date information of further returns On the other hand, extreme high fixed assets growth rate (i.e greater than 1) which could potential source of noise in the analysis

is excluded from the sample

We construct three separated models (1), (2) and (3) to examine the impacts of managerial ownership, block ownership and state ownership to firm performance respectively Average Tobin’s Q (AvQ) is utilized as the proxy of firm performance Since failing to incorporate dynamic relation between dependent and independent variables could lead to the dynamic panel bias problem (Bond, 2002), the

first lag of AvQ (AvQ it-1 ) is included into the right-hand-side variables In addition, the quadratic and

cubic terms of ownership structure are employed to examine the commonly observed non-monotonic relationship The set of control variables in all three models are firm characteristics that possibly affect

firm performance, including Size, Lev, FixAGrR, Beta, Age and year dummies The definition of

variables is provided in Table 1

𝐴𝑣𝑄𝑖𝑡 = 𝛼 + 𝜃𝐴𝑣𝑄𝑖𝑡−1+ 𝛽1𝑀𝑂𝑖𝑡+ 𝛽3𝑀𝑂𝑖𝑡2 + 𝛽3𝑀𝑂𝑖𝑡3 + 𝛽4𝐵𝑂𝑖𝑡+ 𝛽5𝐵𝑂𝑖𝑡2 + 𝛽6𝑁𝑜𝐵𝑂 + 𝛽7𝐿𝑂𝐶ℎ𝑎𝑛𝑔𝑒 +

𝛽8𝐶𝑡𝑟𝑙𝐷𝑢𝑚 + + ∑ 𝛿𝑘𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑖𝑡𝑘

Table 1 Variable Definition

Average

Tobin’s Q

AvQ Average of Tobin’s Q at the beginning and the end of the year

Tobin’s Q is computed as market value of equity plus book value

of debt, all divided by book value of total assets

Managerial

ownership

MO Total fraction of shares owned by Board of Directors, Supervisory

Board and Executive Board State

ownership

SO Total fraction of shares owned by central government, local

government, the SCIC and other SOEs Block

ownership

BO Total fraction of shares owned by blockholders, i.e who own ≥

5% of total shares If company does not have any blockholders,

CtrlDum CtrlDumt = 1 if there is a controlling shareholder in the company,

i.e the shareholder own ≥ 50% total shares) in year t, 0 otherwise

Firm size Size Average of total assets at the beginning and the end of the year Leverage Lev Ratio of total debt to total assets

Growth

Opportunity

FixAGrR Fixed assets growth rate

Market risk Beta Beta coefficient obtained from the CAPM-type regression of

weekly stock returns on market returns Market returns is proxied

by the percentage change of VN-Index

Firm Age Age Numbers of year the firm has been listed on HOSE

Year dummies D2007-D2015 Nine year dummies for years from 2007 to 2011

It is widely accepted that there are at least two sources of endogenous ownership structure, including simultaneity and unobserved heterogeneity Numbers of technique have been developed to control for the problem, such as fixed effects and instrumental variables (IV) While fixed effects estimator could only control for unobserved time-invariants, IV strictly requires external strong instrumental variables, which are normally very hard to find in practice (Himmelberg et al., 1999) Thus, instead of using these traditional methods, this study utilizes the system dynamic generalized method of moments (system-GMM) estimator This method was initially introduced by Holtz-Eakin, Newey, and Rosen (1988) and Arellano and Bond (1991), and further developed by Arellano and Bover (1995) and Blundell and Bond (1998) System-GMM on one hand is able to account for unobserved heterogeneity through a system

of first-differenced and level equations, on the other hand can exploit the use of internally generated instrumental variables Therefore, it can overcome the drawbacks of the IV estimator

Except for firm age and year dummies, which are obviously strictly exogenous, following Wintoki, Linck, and Netter (2012) we treat all other variables as endogenous As the first lag of the dependent variable

is included in the dynamic models, lags of two periods or more could be used as instruments for endogenous variables in the system-GMM Specifically, similar to De Miguel et al (2004), we use lags

2 to 4 of level variables as instruments for the first differenced equation, and lag 1 of differenced variables for the level equation The choice between one-step and two-step GMM should also be considered Although two-step is asymptotically more efficient, it tend to sufferer more severely from the downward bias of standard errors (Arellano & Bond, 1991; Blundell & Bond, 1998) Fortunately, this bias could be corrected by the process developed by Windmeijer (2005) that makes two-step robust become more popular recently Because of these reasons, we employ the two-step robust system GMM estimator in our analysis3

4 Data Analysis

3 We use “xtabond2” package written by Roodman (2009) to run two-step system-GMM in Stata The

collapse option is chosen to limit the number of instruments

4.1 Descriptive Statistics

Table 2 shows the correlation matrix and overall descriptive statistics of variables Firm size (Size) -

measured by average total assets - varies significantly among companies, ranging from 76 to 26,624

billion VND The same trend is observed in fixed assets growth rate (FixAGrR) While its mean is

negligible at 5%, the standard deviation is five times larger Interestingly, beta is relatively small with the mean of 0.25 and all values are smaller than 1.00, reflecting low market risk of the manufacturing sector

Table 2 Correlation Matrix and Descriptive Statistics

Size is measured in billion VND

On the other hand, equity ownership is very concentrated with the average of block ownership and managerial ownership are 15% and 54% respectively However, managerial ownership varies much more than block ownership, reflecting the fact that managers could be either blockholders or minority shareholders Although average state ownership is at moderate level of 20%, it varies significantly among companies (ranging from 0% to 84 percent, with standard deviation of 0.24) In respect to firm performance, the value of 1.15 of Tobin’s Q indicates the manufacturing stocks are slightly overvalued during the research period

More details of block ownership is presented in Table 3 Blockholders are defined as shareholders who owns ≥ 5% of total shares The fraction of shares owned by the first, the second and the third blockholder is fairly stable overtime in terms of all mean, median and standard deviation On average, the largest shareholder owns around 40% of total shares While shares owned by the 2nd and the 3rdblockholders are relatively similar, the gap between the largest ownership and the second is considerable, i.e nearly three times higher In addition, the average number of blockholders is around 2.5 for all years, indicating that ownership of manufacturing companies on the HOSE is extremely concentrated, and the majority of shares is on hand of very few shareholders

Table 3 Breakdown of block ownership

Mean (median) [standard deviation]

2007 2008 2009 2010 2011 2012 2013 2014 2015 Pool

Block

ownership

0.43 0.46 0.48 0.47 0.51 0.55 0.57 0.55 0.56 0.54 (0.35) (0.44) (0.50) (0.46) (0.51) (0.55) (0.56) (0.56) (0.57) (0.55) [0.26] [0.21] [0.16] [0.16] [0.15] [0.14] [0.15] [0.19] [0.19] [0.17]

The largest

ownership

0.21 0.34 0.35 0.35 0.36 0.39 0.40 0.40 0.42 0.38 (0.09) (0.42) (0.39) (0.38) (0.38) (0.40) (0.43) (0.43) (0.45) (0.40) [0.26] [0.19] [0.17] [0.17] [0.16] [0.16] [0.16] [0.19] [0.19] [0.18]

The 2nd

largest

ownership

0.10 0.12 0.11 0.10 0.12 0.13 0.14 0.13 0.14 0.13 (0.07) (0.10) (0.10) (0.09) (0.11) (0.12) (0.14) (0.11) (0.11) (0.11) [0.07] [0.08] [0.05] [0.04] [0.06] [0.06] [0.07] [0.07] [0.08] [0.07]

The 3rd

largest

ownership

0.06 0.07 0.07 0.08 0.09 0.09 0.09 0.10 0.09 0.09 (0.06) (0.07) (0.07) (0.08) (0.09) (0.08) (0.09) (0.08) (0.08) (0.08) 0.01 0.01 0.02 0.03 0.02 0.03 0.03 0.04 0.03 0.03

Numbers of

blockholders

4.00 2.17 2.38 2.41 2.48 2.68 2.70 2.55 2.39 2.54 (3.5) (2) (2) (2) (2) (3) (3) (2) (2) (2) [2.16] [1.03] [1.30] [1.05] [1.38] [1.47] [1.53] [1.53] [1.50] [1.44]

Managerial ownership is further broken down into BOD, Supervisory Board and Executive Board ownership, which is showed in Table 4 All three components of managerial ownership only vary slightly during the research period This very stability could be attributable to the preservation or reappointment

of manager positions – the situation usually being observed in the inactive managerial labor market of

an emerging country as Vietnam

Table 4 Breakdown of managerial ownership

Mean (median) [standard deviation]