Access the full text.
Sign up today, get DeepDyve free for 14 days.
Neri Karra, P. Tracey, N. Phillips (2006)
Altruism and Agency in the Family Firm: Exploring the Role of Family, Kinship, and EthnicityEntrepreneurship Theory and Practice, 30
(2007)
Family firm’s manager behavior choice based on the agency and stewardship theory
James Chrisman, J. Chua, F. Kellermanns, Erick Chang (2007)
Are family managers agents or stewards? An exploratory study in privately held family firmsJournal of Business Research, 60
(2014)
Does altruistic behavior reduce the agency cost? An empirical study based on kin altruistic behavior in family business
Frans Stokman, M. Assen, J. Knoop, Reinier Oosten (2000)
Strategic Decision MakingTransforming Financial Institutions
Victoria Dickinson (2011)
Cash Flow Patterns as a Proxy for Firm Life CycleThe Accounting Review, 86
(2011)
Life cycle stage, corporate governance and capital allocation efficiency
H. Skaife, D. Collins, William Jr., Ryan LaFond (2007)
The Effect of SOX Internal Control Deficiencies and Their Remediation on Accrual QualityLSN: Corporate Governance U.S. (Topic)
Dechun Wang (2006)
Founding Family Ownership and Earnings QualityJournal of Accounting Research, 44
Vikas Mehrotra, R. Morck, J. Shim, Yupana Wiwattanakantang (2011)
Adoptive Expectations: Rising Sons in Japanese Family FirmsOrganizations & Markets: Formal & Informal Structures eJournal
Mikko Mustakallio, E. Autio, S. Zahra (2002)
Relational and Contractual Governance in Family Firms: Effects on Strategic Decision MakingFamily Business Review, 15
H. Skaife, David Veenman, D. Wangerin (2013)
Internal Control Over Financial Reporting and Managerial Rent Extraction: Evidence from the Profitability of Insider TradingJournal of Accounting and Economics, 55
Clifford Smith, Jerold Warner (1979)
On financial contracting: An analysis of bond covenantsJournal of Financial Economics, 7
S. Finkelstein (1992)
Power in top management teams: dimensions, measurement, and validation.Academy of Management journal. Academy of Management, 35 3
吕长江, 赵宜一 (2015)
Family or Money?——— The Impact of Family Relationship on Compensation Contract in Family Firms
A. Leone, M. Liu (2008)
Accounting Irregularities and Executive Turnover in Founder-Managed FirmsEconometrics: Econometric & Statistical Methods - General eJournal
(1947)
Rural China
Henry Mintzberg, J. Waters (1990)
Tracking Strategy in an Entrepreneurial FirmFamily Business Review, 3
R. Drazin, Robert Kazanjian (1990)
RESEARCH NOTES AND COMMUNICATIONS A REANALYSIS OF MILLER AND FRIESEN'S LIFE CYCLE DATA
(2008)
An analysis of high-tech enterprise corporate governance mechanism based on power game: From the corporate governance perspective
(2009)
Family authority and firm value: Empirical evidence from family firms
L. Gómez‐Mejía, R. Wiseman (2007)
Does agency theory have universal relevance? A reply to Lubatkin, Lane, Collin, and VeryJournal of Organizational Behavior, 28
(2016)
How do CEOs see their role? Management philosophies and styles in family and non-family firms
(1989)
How and why corporation grow and die and what to do about it: Corporate life cycle
Robert Bushman, Abbie Smith (2001)
Financial Accounting Information and Corporate GovernanceJournal of Financial Abstracts eJournal
L. Berghe, Steven Carchon (2003)
Agency Relations within the Family Business System: An Exploratory ApproachERPN: Agency Issues (Sub-Topic)
M. Lubatkin, W. Schulze (2007)
An Organizational Justice-Based View of Self-Control and Agency Costs in Family FirmsWiley-Blackwell: Journal of Management Studies
(2010)
An investigation and analysis of the effectiveness of internal control and its influencing factors in China
J. Aharony, H. Falk, Nir Yehuda (2004)
Corporate Life Cycle and the Value Relevance of Cash Flow versus Accrual Financial Information
Mei Feng, Chan Li, Sarah McVay (2009)
Internal Control and Management GuidanceSocial Science Research Network
(2000)
Life-cycle stage effects on the incremental information content of earnings and cash flows
Joseph Fan, T. Wong (2000)
Corporate Ownership Structure and the Informativeness of Accounting Earnings in East AsiaAFA 2001 New Orleans Meetings (Archive)
G. Gong, B. Ke, Yong Yu (2012)
Home Country Investor Protection, Ownership Structure and Cross-listed Firms’ Compliance with SOX-mandated Internal Control Deficiency DisclosuresINTL: Managing in Emerging Markets (Topic)
Ronald Anderson, D. Reeb (2003)
Founding-Family Ownership and Firm Performance: Evidence from the S&P 500Journal of Finance, 58
Ronald Anderson, D. Reeb (2003)
Founding‐Family Ownership, Corporate Diversification, and Firm Leverage*The Journal of Law and Economics, 46
W. Schulze, M. Lubatkin, Richard Dino, Ann Buchholtz (2001)
Agency Relationships in Family Firms: Theory and EvidenceOrganization Science, 12
J. Davis, F. Schoorman, L. Donaldson (1997)
TOWARD A STEWARDSHIP THEORY OF MANAGEMENTAcademy of Management Review, 22
Hu Mingxi (2015)
Property rights,managerial power and internal controlJournal of Chongqing University
S. Zahra, J. Pearce (1989)
Boards of Directors and Corporate Financial Performance: A Review and Integrative ModelJournal of Management, 15
(2011)
The nature of the top managers in clannish firms and the quality of earnings therein
(2009)
The nature and structure of internal control
K. Eisenhardt (1989)
Agency Theory: An Assessment and ReviewAcademy of Management Review, 14
William Mullins, A. Schoar (2013)
How Do CEOs See Their Roles? Management Philosophies and Styles in Family and Non-Family FirmsCorporate Governance: Internal Governance
J. Doyle, Weili Ge, Sarah McVay (2007)
Accruals Quality and Internal Control Over Financial ReportingCorporate Finance: Governance
Pramuan Bunkanwanicha, Joseph Fan, Yupana Wiwattanakantang (2013)
The Value of Marriage to Family FirmsJournal of Financial and Quantitative Analysis, 48
(1947)
Rural China. Shanghai: Sanlian Press (In Chinese)
I. Bardhan, Shu-Ping Lin, Shu-ling Wu (2015)
The Quality of Internal Control over Financial Reporting in Family FirmsAccounting Horizons, 29
Ashiq Ali, Tai-Yuan Chen, S. Radhakrishnan (2006)
Corporate Disclosures by Family FirmsHong Kong: Accounting (Topic)
M. Allen (1981)
Managerial Power and Tenure in the Large CorporationSocial Forces, 60
(2010)
Family member composition and corporate governance efficiency: Empirical research from family firms
D. Hambrick, P. Mason (1984)
Upper Echelons: The Organization as a Reflection of Its Top ManagersAcademy of Management Review, 9
China Journal of aCC ounting StudieS , 2017 Vol . 5, no . 2, 211–233 https://doi.org/10.1080/21697213.2017.1339429 Life cycle stage effects, CEO power and internal control quality: evidence from listed family firms in China* a b Mingxia Hu and Shengdao Gan a uditing d epartment, School of Management, Southwest university of Political Science and law, Chongqing, China; a ccounting d epartment, School of Business, Sichuan university, Chengdu, China ABSTRACT KEYWORDS Ceo power; internal control We study the evolution of internal control quality within listed quality; kinship; life cycle family firms in China, over different stages of the life cycle. We find a stage downward trend in internal control quality, and a significant difference between growth and non-growth stages. Further, we investigate the relationship between CEO power and internal control quality. In the growth stage, the higher the CEO’s structure power and expert power, the higher the internal control quality. In addition, the higher the CEO’s ownership power, the lower the internal control quality. But in the non-growth stage, a CEO’s expert power and ownership power is significantly negatively associated with the internal control quality. Additionally, the different types of CEO source in the family firm result in heterogeneity of CEO power. The findings show that the actual controller as CEO prefers weaker internal control quality and has a positive moderating effect on the association of expert power and the internal control quality. Where the CEO is related as a family member of the actual controller, then the closer is the kinship network and the lower is the internal control quality. Moreover, a closer kinship network increases the negative impact of ownership power on internal control quality. 1. Introduction Since China’s Opening-up reform in 1978, family firms have become an important form of organisation. According to Forbes China Rich List (2014), there were 747 listed family firms out of 1,485 A-share listed private companies on 31 July 2014. Because of the concentrated ownership of the founding family and ‘altruism’ arising out of family kinship in China, family members are active participants in management, which makes the interests of managers and shareholders more aligned. Therefore, in family firms, the agency conflicts between owners and managers are lower while the owner–owner conflicts are higher (Ali, Chen, & Radhakrishnan, 2007; Anderson & Reeb, 2003a, 2003b). Prior literature has focused primarily on characteristics of family firms such as organisational governance, ownership structure and the quality of financial reporting (Ali et al., 2007; Fan & Wong, 2002; Wang, 2006), while CONTACT Mingxia hu 806926637@qq.com, hu.1519@osu.edu Paper accepted by Cong Wang. © 2017 a ccounting Society of China 212 M. HU AND S. GAN ignoring internal control quality. The Fundamental Rules of Enterprise Internal Control explicitly state that the management and chairman of the board are the executive body and the responsible entity respectively. The internal control system restricts potential abusive rights of the executive by segregation of different duties. On the one hand, the CEO is respon - sible for establishing rules and regulations. On the other hand, the CEO has the right to deal with some exceptions. Prior research has focused on the relation of CEO power and internal control quality from the property rights aspect (Hu, Gan, & Lu, 2015). According to the life cycle theory, the formation and development of a corporation have elements of the characteristics of a living body (Adizes, 1989; Drazin & Kazanjian, 1990). The production, operation, organisational characteristics and the relationship between investors and managers are distinct in different life cycle stages, which leads to information asymmetry and agency conflicts. The uniqueness of a family firm is typically characterised by the ‘altru- ism’ from a family member’s kinship ( Van den Berghe & Carchon, 2003). The effects of altruism on agency conflicts have periodic differences (Karra, Tracey, & Phillips, 2006). The essence of internal control is to reduce agency costs through checks and balances (Xie, 2009). Whether the internal control quality has life cycle stage characteristics and the relationship of CEO power and internal control quality has dynamic changes, are topics that merit investigation, in order to extend prior literature. Using a sample of Shanghai and Shenzhen A-share listed family firms from 2007 to 2013, we investigate the dynamic evolution of internal control quality in listed family firms and the effects of CEO power in a family firm on the internal control quality at different stages of the life cycle. Furthermore, we examine the moderating effects of different types of CEO on the relationship of CEO power and the internal control quality. Our empirical evidence shows a downward trend in the internal control quality in a family firm between growth and non-growth stages. In the growth stage, the power structure and expert power of the CEO (as evidenced by the period of CEO tenure) are positively related with internal control quality, and the ownership power of the CEO is negatively correlated with internal control quality. In the non-growth stage, the expert power of the CEO and the ownership power have a negative correlation with internal control quality. We further collect background information on the CEO manually and categorise the CEO types as follows: actual controller as CEO; family member of actual controller as CEO; co-founder outside the family as CEO; and non-founder non-family member as CEO. Among those types of CEO, the family member of the actual controller as CEO can be further divided into the actual controller’s spouse, parents and children, siblings, son-in-law, and other rel- atives. Non-family member non-founders as CEO can be categorised as internal promotion of employees and external professional managers. We find the actual controller as CEO has a direct negative impact on internal control quality and has a positive moderating effect on the relation of the expert power of the CEO and internal control quality. The theory of ‘the Pattern of Differential Sequence’ (Fei, 1947), claims: ‘the kinship network in China’s society starts with a self-centred circle, and the intimacy of the relationship is descending as the network enlarges like a stone thrown into water, whose ripples formed f undamental rules of enterprise i nternal Control were issued by China Ministries in 2008, requiring listed firms in China to self-evaluate the effectiveness of internal control system, disclose their annual self-evaluation report, and hire accounting firms to issue an audit opinion on the effectiveness of the internal control system. CHINA JOURNAL OF ACCOUNTING STUDIES 213 become larger and farther’. The main force of driving the ripple is kinship, such as ‘parent-chil- dren and siblings’. According to the survey on the importance of family member’s kinship network conducted by He and Lian (2009), we categorise the family member CEO based on kinship network from the actual controller and find that the closer the kinship network, the lower the internal control quality. In addition, the closer kinship network can increase the negative effect of CEO’s ownership power on the internal control quality. The research on these issues contributes to a comprehensive understanding of the evo- lution of internal control quality in listed firms in China at different stages of the life cycle. It can also be beneficial to recognise the effect of CEO power on the internal control quality dynamically. It has a profound policy implication on a China-listed family firm’s internal control system adjustment and reform in distinctive life cycle stages. The contributions of this study of family firms are three-fold. First, while prior studies have focused on the measurement of firm value, financial performance, earnings quality and voluntary disclosure to investigate agency conflicts within family firms, we study the dynamic evolution of internal control quality in listed family firms in China. Our study complements the existing stream of research by analysing the internal control quality characteristics in different life cycle stages, which enriches the literature on internal control. Second, our study investigates the relationship between CEO power in different dimen - sions and internal control quality dynamically, which provides empirical evidence on the evolution of the internal control system of listed family firms in China. We find the CEO’s higher structure power and expert power are likely to be associated with a stronger internal control quality in the growth stage. Our study also contributes to the literature on internal control by identifying CEO power in a family firm in China as a determinant of a firm’s internal control quality. Hence, the research complements Skaife, Collins, Kinney, and LaFond (2007) and Doyle, Ge, and McVay (2007), who propose and examine general firm attributes and industry characteristics as determinants of material weakness in internal control. Third, we categorise CEO types according to unique characteristics of kinship network, entrepreneurship and other important factors in family firms in China and we verify the moderating effect of the type of CEO. In particular, our findings show that where there is a family member of the actual controller as CEO then there is a relative preference for weak internal control quality in order to acquire private benefits. Our evidence shows kinship is used by controlling shareholders to entrench themselves in the governance of family firms. The structure of the paper is as follows. The second section provides the theoretical back- ground and hypothesis of the research. The third section illustrates the research design, and the fourth section reports and discusses the main findings of the empirical study. In the fifth section, we adopt multi-methods to mitigate the potential endogeneity between CEO power and internal control quality. Finally, we summarise our conclusions. Kinship network is defined as an order of the kinship between two individuals. t he kinship includes the nuclear family and relatives. t he nuclear family includes spouse, parents and children, and siblings. t he order of ‘spouse > parents and chil- dren > siblings’ is applied in the nuclear family. relatives include son-in-law, daughter-in-law and other distant relatives. We assume the order of spouse, parents and children, siblings, son-in-law, other distant relatives, and non-family members. 214 M. HU AND S. GAN 2. Theory background and hypothesis development 2.1. Life cycle stage effects and internal control quality In prior research, scholars have introduced life cycle theory to accounting research. Black (2000) studies the relationship between earnings as well as cash flow information and life cycle, and concludes cash flow includes more information than earnings in the stages of growth, maturity and decline. Aharony, Falk, and Yehuda (2004) show that the explanatory power of cash flow is weaker than that of accounting information under the accrual basis using life cycle theory. Li, Li, and Tang (2011) apply life cycle theory to explain the dynamic changes of capital allocation efficiency and the dynamic mechanism of corporate govern- ance. With regard to internal control quality, Zhang and Zheng (2010), using a questionnaire survey, conclude there is higher internal control quality when firms are in the mature stage. According to the corporation life cycle theory, a corporation’s formation and development have similarities to living organism. Corporations have different characteristics of organisa- tional features, operations and management decisions as well as different agency costs in their distinctive developing stages. Agency theory explains why agents may not act in the interests of the principals unless their behaviour is closely monitored (Eisenhardt, 1989). Financial reporting systems play an important role in alleviating agency conflicts. Financial information can be used to directly evaluate manager stewardship and as contractual inputs to restrict manager’s opportunistic behaviour (Bushman & Smith, 2001; Smith & Warner, 1979). For the financial reporting system to fulfil its role, it is essential to have higher internal control quality. The internal control system is an important line of defence against financial reporting misstatements. The weaker the internal control quality, the poorer the accounting quality and the less dissemination of reliable financial information to outside investors (Feng, Li, & McVay, 2009). For listed family firms in China, the agency conflicts are not obvious in the start-up stage because of the small size and the founder acting as owner. It leads to a lower demand for an internal control system. With constant expansion of organisation size in the growth stage, ownership is gradually segregated from management rights, the organ- isational hierarchy becomes more sophisticated, and agency problems begin to emerge. The most outstanding characteristics of agency conflicts between founder and family man- ager are embedded in ‘altruistic behaviour’, which has a periodic effect on the agency costs (Wang, Xu, & Wang, 2014). In the corporation’s early growth stage, with the organisation growth, the positive effect of altruistic behaviour gradually decreases and the negative impact rises. First, in the growth stage, the family members are involved in a high proportion of busi- ness management. The dual trust symmetry between family members and founders is helpful to reduce the agency costs between managers and owners (Su, 2007). The founders can clearly know the effort spent at work and effectively eliminate opportunistic behaviour by a family manager (Lubatkin, Ling, & Schulze, 2007). Second, due to kinship, family members have residual allocation rights to the wealth of the corporation. The interests of owners and managers are naturally tied by kinship, which may reduce the moral hazard. Kinship can mitigate the information asymmetry between founders and managers, which reduces the owner’s monitoring costs (Schulze, Lubatkin, Dino, & Buchholtz, 2001). ‘Mutual social capital and trust’ in the family ties lead to the mutual interests among family members. Driven by the mutual interests, family managers tend to implement altruistic behaviour. Under the value of altruism, an individual may consider the interests of others. Family members play CHINA JOURNAL OF ACCOUNTING STUDIES 215 roles of ‘stewardship’ in the family wealth to maximise family interests (Davis, Schoorman, & Donaldson, 1997). Internal control system aims to improve the quality of financial reporting information, enhance the operating effectiveness and efficiency of a corporation and achieve its developing strategy. Generally speaking, family managers in the growth stage prefer high internal control quality for the firm’s long-term interest due to the altruistic positive effect. Hence, we argue family firms may have relatively higher internal control quality in the growth stage. As family firms continue to grow and expand, the positive effect of ‘altruism’ gradually declines while the negative effect rises. First, ownership will be further decentralised. Family members possess less, or even no ownership. External professional managers are gradually introduced. The moral hazard of family managers is increased due to limitations in their knowledge and skill. On the other hand, with the continuous development and expansion of firms, the number of agent chains increases and as the single agent chain is lengthened, the amount of information expands dramatically. Children, as members of the nuclear family, regroup into another family after reaching a certain age. Accordingly, the degree of infor- mation asymmetry is deepened. At this time, founders lack sufficient information to know the family manager’s effort on work, and also it is more difficult to punish their opportunistic behaviour. Meanwhile, distant family members and various non-nuclear family members are introduced into the firm’s management, which aggravates the adverse selection problem and worsens multiple discrimination issues among different type of family members within the pattern of differential sequence (He, Li, & Chen, 2010). Therefore, the altruism has increas - ingly negative influences on agency costs, and the opportunistic behaviour of family man- agers also increases. Family managers may not be motivated to invest a lot of resources to construct an internal control system just to conceal opportunistic behaviour, or the internal control system may even become part of the agency problem. Second, the family firm’s agency conflicts are more complex and severe, including the conflicts between family man - ager, external professional managers and controlling shareholders (Su, 2007). Different inter - est groups are formed in the family firms, and the common sense of identity is lowered within family members and external professional managers. Therefore, the transaction and coor- dination costs of information communication and interaction may increase, which further reduces the efficiency of an internal control system. In the decline stage, the growth oppor - tunities shrivel, organisational flexibility declines, and internal bureaucracy increases. When the family firms are faced with the risk of being merged, they are inclined to entrench them - selves. For the purpose of entrenchment, the family owners may prefer to establish a weaker internal control system in order to extract private rent, because the lower quality of an internal control system leads to poor accounting quality and the less reliable dissemination of financial information to outside investors, which aggravates the information asymmetry between the controlling and minority shareholders (Bardhan, Lin, & Wu, 2015). It is also possible they may circumvent the existing internal control system and foster a weaker inter- nal control environment in the maturity and decline stages. For the above reasons, we posit the following hypotheses: H1-1: For listed family firms in China in the growth stage, the internal control quality is signif- icantly higher. H1-2: For listed family firms in China in the non-growth stage, the internal control quality is significantly lower. 216 M. HU AND S. GAN 2.2. Life cycle effects, CEO power and internal control quality When corporations are at different development stages, CEO power has different character - istics. Ma, Li, and He (2008) state that the fundamental starting point of life cycle effects on management decision is attributed to the power allocation. Therefore, if the life cycle effects are neglected, it is difficult to investigate the effects of CEO power on the internal control quality. CEO power can be classified as structure power, expert power, ownership power and reputation power, as proposed by Finkelstein (1992). Because the reputation mechanism of professional managers has not been fully established in the process of China’s economic transition, we use structure power, expert power and ownership power to measure the CEO power. According to the ‘upper echelons theory’, in which the organisation is a reflection of the personal characteristics of top management at large, the power structure has an enor- mous impact on the efficiency of implementation of organisational policy (Hambrick and Mason, 1984). First, from the perspective of CEO ownership power, in the growth stage, since the CEO is most likely to be the founder or a member of the founder’s family, the CEO plays a role as a family owner, which is the source of CEO ownership power. The high proportion of CEOs who are family owners and the related altruism by family kinship can reduce the conflicts between manager and shareholder, but also aggravate agency conflicts between the con- trolling and minority shareholders. The internal control system limits the CEO in exploiting private interests through restricting false information disclosure, restricting manipulation of accounting information and restricting insider trading. The family CEO is able to exploit shareholders by means such as inflated compensation, special dividends, tunnelling, or self-dealing (Leone & Liu, 2010). Skaife, Veenman, & Wangerin (2013) find the likelihood of insider trading is greater among firms that have more internal control weakness compared with firms with high internal control quality. The weaker internal control quality impairs the ability of minority shareholders and other contractual participants to intervene in the CEO’s opportunistic behaviour. When the CEO possesses more concentrated ownership, which often allows their family members to hold positions in top management or their firm’s boards (Mullins & Scholar, 2016), family owners as CEO have more capacity to mitigate the moni- toring efficiency of the board of directors. In this circumstance, the internal control quality may be lower. From the perspective of the governance mechanism of family firms, the rela- tionship of family shareholders and family managers is based on implicit contracts, such as trust and emotion, to reduce the agency conflicts (Chrisman, Chua, Kellermanns, & Chang, 2007). Family firms rely on more relationship control, including loyalty, altruism and blood relations, which complement or substitute the formal contractual control (Mustakallio, Autio, & Zahra, 2002). With the increase in CEO ownership power, the family shareholders may rely on more informal governance, which allows more formal control to be replaced or lacks adequate resource to input, which is harmful for the family firm’s internal control construc - tion. Hence, we argue that the higher is the CEO ownership power, the lower is the internal control quality. In the growth stage, because of the positive effect of altruism, CEOs may have incentives to maximise the corporation’s interests. From the perspective of CEO expert power, with the increase of CEO tenure, a CEO can accumulate higher personal prestige, which facilitates forming CEO authority. CEO authority is advantageous to coordinate family members’ inter- ests and organisational objectives. An authoritarian culture possesses the efficiency advan- tage of internal control design soundness. We argue long-term attributes provide compelling CHINA JOURNAL OF ACCOUNTING STUDIES 217 economic incentives to pursue high internal control quality that is essential to a firm’s long- term survival and success. In this situation, the higher CEO expert power is beneficial to the formation of system authority, which can generate obvious efficiency. With the CEO’s longer tenure in the firms, the CEO can acquire superior knowledge about the business, which enables the CEO to better monitor internal controls, detect control problems and rectify controls accordingly. Hence, we argue the higher the CEO expert power, the higher the internal control quality. From the perspective of structure power, CEO duality and a board of directors controlled by internal directors can motivate the high internal control quality in the growth stage. In the growth stage, the family CEO is largely involved in the business, the CEO duality helps to establish unity of command and clarifies decision-making authority, which can reduce or eliminate a power struggle within the family firms, improve the operation efficiency, and reduce organisational and coordination problems caused by segregation of duties (Mintzberg & Waters, 1990). It is helpful to increase the efficiency of implementation of internal control. On the other hand, in the information communication process, CEO duality results in extreme ideas during the decision-making process (Zahra & Pearce, 1989). In this circumstance, it is easier to form an authoritarian culture. In this situation, the internal control system may be implemented in a top-down way, which makes the implementation of internal control more efficient. So we argue the CEO structure power is positively related with the internal control quality. Based the above analysis, we posit the following hypotheses. H2-1: For listed family firms in China in the growth stage, CEO ownership power is negatively related with the internal control quality. H2-2: For listed family firms in China in the growth stage, CEO expert power is positively related with the internal control quality. H2-3: For listed family firms in China in the growth stage, CEO structure power is positively related with the internal control quality. With the expansion of corporation size and an increasingly competitive market, the external professional managers are introduced, which makes the power allocation between profes- sional managers and family members more complex. The agency conflicts include founders and family managers, founders and professional managers, controlling and minority share- holders, which lead to complex agency problems. As family firms accelerate the introduction of closer and more remote relatives as family managers, all kinds of non-nuclear family members intensify the adverse selection problem. Under these circumstances, the agency heterogeneity in family firms is intensified, which may easily lead to indifferent shared values and weak internal cohesion power. On the other hand, the agency characteristics increase the family manager’s opportunistic behaviour. An internal control system is an important line of defence to protect the outside investors and reduce agency conflicts. As the CEO ownership power increases, they have more incentive and capacity to implement a weaker internal control system. Therefore, we posit that the higher the CEO ownership power, the weaker the internal control quality. From the perspective of CEO expert power in the non-growth stages, CEOs may prefer to entrench themselves as agents, which hampers the internal control quality. Allen (1981) argues that with the longer tenure in the firm’s position, the CEO’s knowledge, social expe - rience and business capacity will be greatly improved, which means that the firm has a CEO possessing more internal and external information about the firm. Without rigorous 218 M. HU AND S. GAN monitoring of the board of directors, CEOs can entrench themselves more successfully. The internal control system can reduce the ability of the CEO to acquire a private interest through fake information disclosure, accounting information manipulation or insider trading. The high internal control system quality can prevent the outsider’s intervention and monitoring. Hence, as the CEO expert power increases, he or she may have more power to weaken the internal control system to obtain private rent. From the perspective of CEO structure power, duality compromises the ability of the board to reasonably monitor a CEO’s practices, policies and performance. If the CEO serves as chairman, he or she can easily mitigate the monitoring function of a board of directors and audit committee, which results in weaker internal control quality. Moreover, CEO duality means that the decision rights and ownership rights overlap, so that a check on power is not realised effectively. In addition, compared with the involvement of many executives in the decision-making process, CEO duality gives them more decision-making freedom, and allows their own preference to be dominant (Eisenhardt & Zbaracki, 1992), which hampers effective information communication in the implementation of an internal control system. Based on the analysis above, we posit the following hypotheses. H3-1: For listed family firms in China in the growth stage, CEO ownership power is negatively related with the internal control quality. H3-2: For listed family firms in China in the growth stage, CEO expert power is negatively related with the internal control quality. H3-3: For listed family firms in China in the growth stage, CEO structure power is negatively related with the internal control quality. 3. Research design 3.1. Sampling 3.1.1. Determination of family firms and sample selection Based on the CCER database, A-share listed family firms from 2007 to 2013 were selected. With reference to the approach of Xu and Lv (2011), family firms were defined as follows: actual controller or natural person, or private firms embedded with blood and marriage. We eliminated the following: (1) ST (Special Treatment) firms; (2) finance and insurance firms; (3) firms for which the required financial data are unavailable in CCER or CSMAR. One percent and 99% financial variable data were curtailed. We got a sample of 3,635 firm-year observa- tions. Table 1 reports the sample distribution by year and industry. The data are from CSMAR and CCER databases, the CEO type and CEO tenure are acquired manually. Internal control indexes are from Dibo Risk Management Limited Enterprise from 2007 to 2013. t he CCer database is the China e conomic and f inancial database, which is a high standard database product developed by Beijing Sinofin information Services and the China Center for economic research, specialising in academic research in the economic and financial fields. t he listed family firms can be filtered out from the basic information in the Private listed Company data, in which the privatisation approach index includes family listed firms and other private listed firms. St firms (Special t reatment) refer to firms that will face imminent danger of delisting unless they return to profitability after reporting two consecutive annual losses. dibo r isk Management limited enterprise established in 2001, is the first professional organisation with internal control, risk management and internal audit as the main business. it is mainly for large and medium size state-owned enterprises, listed companies, administrative institutions and other large and medium size companies to provide internal control and risk management consulting and software services. CHINA JOURNAL OF ACCOUNTING STUDIES 219 Table 1. Sample distribution by year and industry. Industry code and name 2007 2008 2009 2010 2011 2012 2013 Total a a griculture, forestry, animal husbandry and 6 7 8 9 15 16 17 78 fisheries B Mining 3 3 4 4 4 7 12 37 C Manufacture 129 185 220 272 494 721 773 2,794 d electricity, g as and Water 2 1 1 1 2 3 6 16 e Construction 6 8 9 11 17 25 28 104 f Wholesale 4 5 6 6 9 10 13 53 f retailing 3 3 5 6 9 16 19 61 g t ransportation 2 2 3 3 5 6 5 26 i Software 12 15 18 26 47 67 90 275 K real estate 6 12 11 14 15 13 20 91 l Business Services 0 2 3 3 5 6 6 25 M Professional Services 0 0 0 3 6 7 10 26 n e cological Protection and environmental 0 0 0 0 0 0 7 7 Management r Media and Culture 1 1 1 2 5 7 9 26 S Miscellaneous 2 2 2 2 2 2 4 16 t otal 176 246 291 362 635 906 1,019 3,635 3.1.2. Determination of family firm’s CEO type The classification of the CEO type of a family firm is according to whether the CEO is the founder, and whether there exists a kinship network. The founder CEO includes the actual controller and co-founder even if outside the family. A non-founder CEO includes a family CEO and non-family CEO. A non-family CEO is classified into internal promoted employees and external professional managers. Family CEO can be further sub-divided according to the following rules. (1) Nuclear family members (spouse, parents and children, siblings) are closer than other relatives (Fei, 1947). (2) Based on the research from He and Lian (2009), the order of ‘spouse > parents and children > siblings’ is applied within the nuclear family. (3) Relatives outside the nuclear family, including son-in-law, daughter-in-law and other dis- tant relatives. Bunkanwanicha, Fan, and Wiwattanakantang (2013) and Mehrotra, Morck, Shim, and Wiwattanakantang (2013) reiterate that authority owned by a son/daughter- in-law can enhance a corporation’s value. Therefore, ‘son/daughter-in-law > other distant relatives’ is given. Other distant relatives include cousins, cousins’ spouses, the spouse’s siblings, and other relatives. The number of each type of sample is small and it is diffi- cult to judge the relationship, therefore we group them together. Thus, CEO is sorted based on relationship: actual controller’s spouse, parents and children, siblings, son/ daughter-in-law, other distant relatives, and non-family members (Zhao & Lv, 2015). f irst, we can get the Ceo and actual controller’s name from the listed family firm’s annual financial report. t hen, the rela- tionship between the Ceo and actual controller can be acquired through the publicly disclosed report or network according to the names. in some firms’ publicly disclosed report, such as the annual financial report, board meeting resolution announcement and others, they disclose the Ceo and the actual controllers’ relationship. f or example, for the company (601700), we got the Ceo and actual controller’s name from the annual financial report in 2012. in the company’s financial report in 2012, the Ceo name was liyi f an, the actual controller’s name was Jianggang f an. t he financial report in 2012 disclosed Jiangang f an and liyi f an were in the father-and-son relationship. f or the company (002515), the firm f an. t he financial report in 2012 disclosed Jiangang fChanghuang Xue and the actual controller’s name was Yanjun Shi. i n the firm’s board meeting resolution announcement in 2010, it disclosed the Changhuang Xue and Yanjun Shi were brothers-in-law. and for some firms, we can get the relation between Ceo and the actual controller through published information in the financial media, which directly discloses the relationship. 220 M. HU AND S. GAN 3.2. Empirical model specification and variables 3.2.1. Life cycle stage effects and internal control quality To verify the hypotheses H1-1 and H1-2, we use an OLS regression Equation (1) to directly estimate a corporation’s life cycle stage effects on internal control quality. ln(ICINDEX) = a + a LIFE_CYCLE + a POWER1 i,t 0 1 i,t 2 i,t + a POWER2 + a POWER3 + a Controls + (1) 3 i,t 4 i,t 5 i,t i=1 In Equation (1), the dependent variable ln(ICINDEX) stands for the internal control quality. LIFE_CYCLE consists of growth, maturity and decline stages. CEO power includes structure power (POWER1), expert power (POWER2) and ownership power (POWER3). In the regression process, under the controlling and non-controlling CEO power variables, we study the ee ff cts of life cycle stage on internal control quality. Control variables include LEVERAGE, SIZE, ROA, LARGEST_SHARE, TOTAL_ASSET_GROWTH, INVENTORY_RATIO and BIG4. See Table 2 for defi- nitions and measurements. 3.2.2. CEO power and internal control quality in distinctive life cycle stages For the hypotheses H2-1, H2-2, H2-3, H3-1, H3-2 and H3-3, according to different stages of a corporation’s life cycle, the sample is divided into growth and non-growth sub-samples. Equation (2) is used to examine whether CEO power has presented different life cycle stage effects on the internal control quality. The control variables include LEVERAGE, SIZE, ROA, LARGEST_SHARE, REVENUE_GROWTH, INVENTORY_RATIO, AGE and BIG4. ln(ICINDEX) = a + a POWER1 + a POWER2 + a POWER3 + a Controls + (2) i,t 0 2 i,t 3 i,t 4 i,t 5 i,t i=1 3.2.3. CEO type moderating effect on CEO power and internal control quality Due to the complexity of CEO sources, there may be family members, external professional managers or internal promoted employees who connect with the firm based on economic interests and emotional investment (Gomez-Mejia & Wiseman, 2007). Therefore, it leads to the formation of CEO power heterogeneity. We analyse the relationship of CEO power and internal control quality according to the CEO types and clarify their dominant factors on the internal control quality. Table 2. Control variables definition and measurement. Variable Definition and measurement Data source LEVERAGE t he ratio of asset to liability CSMar SIZE t he natural logarithm of total assets CSMar ROA t he ratio of return on assets CSMar LARGEST_SHARE t he percentage of largest shareholder’s ownership CSMar TOTAL_ASSET_GROWTH t he percentage of current year’s total asset growth CSMar INVENTORY_RATIO t he ratio of inventory to assets CSMar REVENUE_GROWTH t he ratio of revenue growth compared to the previous year CSMar BIG4 if audited by big4, equals 1, otherwise 0 CSMar AGE t he time span of the company going public CSMar CHINA JOURNAL OF ACCOUNTING STUDIES 221 Based on whether the CEO is an entrepreneur, family relationship and the kinship network, the CEO is classified as an actual controller as CEO, co-founder outside the family as CEO, family member of actual controller as CEO, internal promoted employees and external pro- fessional managers as CEO. Equation (3) is used to examine different CEO types’ moderating effects on CEO power and internal control quality. (3) ln(ICINDEX) = a + a POWER1 + a POWER2 + a POWER3 i,t 0 1 i,t 2 i,t 3 i,t + a X {SELF, FCEOTYPE} + a X {SELF, FCEOTYPE} 4 i,t 5 i,t × POWER1 + a X {SELF, FCEOTYPE} × POWER2 i,t 6 i,t i,t + a X {SELF, FCEOTYPE} × POWER3 7 i,t i,t + a Controls + 8 i,t i=1 When X is taken as SELF, it is a dummy variable. For SELF to be 1 means actual controller as CEO, while for SELF to be 0 means others as CEO. When X represents FCEOTYPE, X is an ordinal variable. Family member as CEO is classified as actual controller’s parents and children, sib - lings, son-in-law and other distant relatives based on the kinship network, as FCEOTYPE taking as 4, 3, 2 and 1 successively. The coefficient of FCEOTYPE ×POWER represents the i,t i,t kinship network moderating effect on CEO power and internal control quality. 3.3. Variable measurement and specification 3.3.1. Measurement of internal control quality The internal control index issued by the Dibo Risk Management Limited Enterprise is adopted. The index applies a corporation internal control system as its standard basis, based on the realisation of control compliance, report reliability, assets safety, operation and strategy. The ‘Dibo·China Listed Company Internal Control Index’ uses 1,000 points with the distribution of scores ranging from 0 to 1,000. Internal control quality is measured by the natural loga- rithm of the internal control index. The higher the score, the higher the quality of the internal control. 3.3.2. Life cycle stage measurement We use the cash flow model (Dickinson, 2011) to represent the interaction function between corporation resources, funds, operation capability, external macro-environment and corpo- ration strategy. The combination of cash flow information is applied to divide the entire corporation life cycle. Taking into account the contents and sample size in this study, we combine the introduction and increasing stage as the growth stage, and the phase-out and recession stage are grouped as the decline stage. Therefore, the life cycle stages are classified as growth, maturity and decline stages. The specific classification is as shown in Table 3. 3.3.3. CEO power measurement A power indicator is classified as follows when considering the special background in China. Structure power (POWER1), is measured by CEO duality and the number of internal directors on the board of directors. The higher is POWER1, the weaker is the monitoring function of the board of directors. Expert power (POWER2), is defined as CEO tenure. In general, the 222 M. HU AND S. GAN Table 3. life cycle stage classification. Introduction Growth Maturity Phase-out Phase-out Phase-out Recession Recession operating – + + – + + – – cash flow investing – – – – + + + + cash flow f inancing + + – – + – + – cash flow Table 4. Ceo power indicator definition and measurement. Power dimension Description and measurement Structure power (POWER1) X1: Ceo acted as president or vice president of the board of directors, equals to 2, otherwise 1 X2: t he number of board of directors minus the number of independent directors POWER1: if the sum of X1 and X2 is greater than industry average amount, equals to 1, otherwise 0 expert power (POWER2) X3: t he tenure of Ceo in the position, counted by months POWER2: X3 greater than industry average amount, equals to 1, otherwise 0 o wnership power (POWER3) X4: t he Ceo ownership portion greater than industry average amount, equals to 1, otherwise 0 X5: t he ownership portion of institutional investors lower than industry average amount, equals to 1, otherwise 0 POWER3 = X4 + X5 POWER_EW POWER_EW = (X1 + X2 + X3 + X4 + X5)/5 longer the CEO tenure, the more experienced or higher the prestige, and therefore the CEO has stronger power to control the firm. Ownership power (POWER3), is measured by the portion of CEO ownership and institutional investor’s ownership. The larger the portion of CEO ownership power, the greater the influence on the selection of the board of directors. Since institutional investors are mainly used to externally supervise management deci- sion-making, they can largely restrict management. The lower the proportion of institutional investors’ ownership, the weaker the external supervision over the CEO. A comprehensive indicator (POWER_EW) can be computed by averaging the five indicators above. The specific measurements are illustrated in Table 4. 3.3.4. All control variables measurement Control variables include LEVERAGE, INVENTORY_RATIO, SIZE, ROA, LARGEST_SHARE, AGE, TOTAL_ASSET_GROWTH, REVENUE_GROWTH, BIG4, INDUSTRY and YEAR dummy variables. Table 2 illustrates the definition and measurement. 4. Empirical tests and analysis of results 4.1. Descriptive statistics and Pearson correlation analysis Table 5 presents the descriptive statistics of the variables. The mean value of ln(ICINDEX) is 6.51, ranging from 5.24 to 6.87. The mean value of LARGEST_SHARE is 0.34, the maximum value reaches 0.86, which indicates concentrated ownership. The mean value of LEVERAGE is up to 0.35. As for the Pearson correlation among variables used in this study, non-tabulated results show the correlation coefficient between variables is less than 0.5, which suggests there is no serious multicollinearity in these models. CHINA JOURNAL OF ACCOUNTING STUDIES 223 Table 5. d escriptive statistics. Variables N Mean SD Max Min ln(ICINDEX) 3,635 6.51 0.11 6.87 5.24 POWER1 3,635 0.52 0.49 1.00 0.00 POWER2 3,635 0.47 0.49 1.00 0.00 POWER3 3,635 1.14 0.58 2.00 0.00 LEVERAGE 3,635 0.35 0.19 0.95 0.01 LARGEST_SHARE 3,635 0.34 13.94 0.86 0.19 ROA 3,635 0.05 0.06 0.39 −1.15 TOTAL_ASSET_GROWTH 3,635 0.22 1.29 74.84 −0.53 AGE 3,635 4.55 3.98 24.00 1.00 SIZE 3,635 21.39 0.92 25.51 18.34 INVENTORY_RATIO 3,635 0.15 0.12 0.94 0.00 Table 6 provides the descriptive statistics of the classification of CEO type. Panel A shows the classification of CEO type. The mean value of ln(ICINDEX) is highest in the internal pro - moted employees as CEO samples, which implies the highest internal control quality. In the sample of actual controller as CEO, the mean values of POWER1, POWER2 and POWER3 are highest. Panel B shows the classification of family member as CEO based on the kinship network. In the 275 samples, the number of samples of actual controller’s parents and chil- dren as CEO is 133, siblings as CEO is 66, other relatives as CEO total 61 and son-in-law is 15. In the group of actual controller’s parents and children as CEO, the mean value of ln(ICINDEX) is highest. In the 15 son-in-law samples, the mean values of POWER2 and TOTAL_ASSET_ GROWTH are the largest. Panel C shows the classification of samples in different life cycle stages. In the 3,635 total samples, the proportion accounts for 77% in the growth stage, 17.7% in the maturity stage and 5.3% in the decline stage, which indicates most of China’s listed family firms are still in the growth stage. In the growth stage samples, the percentage of actual controller as CEO reaches 33%, the portion of external professional managers and internal promoted employees as CEO is respectively 22% and 12%. In the maturity stage, the proportion of actual controller as CEO accounts for 6.2%, external professional managers as CEO is 5.36% and internal promoted employees as CEO is 3.63%. There are only 194 samples in the decline stage, and the characteristics of China-listed family firms in the maturity stage are similar to those in the decline stage. Therefore, we consolidate these samples in the two stages as a non-growth stage. In the following section, samples are divided into growth and non-growth stages, and a non-parametric test is con- ducted between these two groups. The results are shown in Table 7. The results show ln(IC- INDEX) presents a downward trend, and the difference is significant between the two stages. POWER2 increases and is significantly different between the two stages. 4.2. Multivariate tests of hypothesis H1 In Table 8, columns (1) to (4) show the empirical results of life cycle stage and internal control quality. Columns (1) and (2) show that the coefficient is significantly positive in the growth stage in controlling and non-controlling power variables (0.012 with t = 2.02, 0.012 with t = 1.96). Columns (3) and (4) indicate it is significantly negative in the non-growth stage (–0.012 with t = –2.02, –0.012 with t = –1.96). Combined with the results in Table 7, we con- clude the internal control quality in the growth stage is higher, while the internal control quality is declining in the non-growth stage, which verifies H1-1 and H1-2. 224 M. HU AND S. GAN Table 6. Ceo type classification descriptive statistics. Panel A: Classification of CEO type non-family member non-founder as Ceo (nfCeo ) internal external f amily members Co-founder out- promoted professional a ctual controller of actual control- side the family employees managers Ceo types as Ceo (Self) ler as Ceo (fCeo ) as Ceo N = 3,635 1,478 275 231 591 1,060 ln(ICINDEX) 6.52 6.52 6.52 6.53 6.51 POWER1 0.73 0.35 0.36 0.30 0.42 POWER2 0.56 0.40 0.43 0.37 0.42 POWER3 1.43 0.98 1.18 0.85 0.93 LARGEST_SHARE 0.35 0.35 0.30 0.33 0.35 LEVERAGE 0.31 0.37 0.39 0.40 0.36 ROA 0.05 0.04 0.05 0.05 0.04 AGE 3.30 5.32 4.10 5.99 5.37 TOTAL_ASSET_ 0.19 0.15 0.18 0.24 0.27 GROWTH SIZE 21.16 21.44 21.37 21.54 21.44 INVENTORY_RATIO 0.13 0.19 0.16 0.18 0.16 Panel B: Classification of family member as CEO based on the kinship network Parents and children (4) Siblings (3) Son-in-law (2) o ther relatives (1) N = 275 133 66 15 61 ln(ICINDEX) 6.52 6.51 6.51 6.51 POWER1 0.25 0.29 0.13 0.37 POWER2 0.24 0.13 0.33 0.24 POWER3 1.14 0.59 0.60 0.29 LARGEST_SHARE 0.35 1.12 0.87 1.10 LEVERAGE 0.36 0.35 0.39 0.39 ROA 0.04 0.04 0.03 0.03 TOTAL_ASSET_GROWTH 0.16 0.11 0.36 0.12 SIZE 21.56 21.26 21.32 21.39 INVENTORY_RATIO 0.17 0.23 0.16 0.21 Panel C: Classification of CEO type based on the life cycle stages non-family member non-founder as Ceo (nfCeo ) f amily members Co-founder internal external a ctual controller of actual control- outside the family promoted professional Ceo t ypes as Ceo (Self) ler as Ceo (fCeo ) as Ceo employees managers N = 3,635 1,478 275 231 591 1,060 growth stage 1,203 210 182 419 783 Maturity stage 225 54 38 132 195 d ecline stage 50 11 11 40 82 Table 7. t est of difference in variables. Classification based on corporation development stage T value growth to Variables Growth Non-growth non-growth ln(ICINDEX) 6.54 6.51 3.95*** POWER1 0.40 0.41 0.07 POWER2 0.20 0.38 −5.98*** POWER3 0.96 0.91 1.31 LEVERAGE 0.46 0.40 4.78*** SIZE 2.16 2.13 5.35*** ROA 0.05 0.06 −0.99 TOTAL_ASSET_GROWTH 0.26 0.10 7.38*** INVENTORY_RATIO 0.17 0.17 0.18 LARGEST_SHARE 0.31 0.30 1.81* note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. CHINA JOURNAL OF ACCOUNTING STUDIES 225 Table 8. regression results of life cycle stage effects and internal control quality. (3) Non-growth (4) Non-growth (1) Growth stage (2) Growth stage stage stage Coefficient Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) (t-value) GROWTH 0.012** 0.012* (2.02) (1.96) NON-GROWTH –0.012** –0.012* (–2.02) (–1.96) POWER1 0.008 0.008 (1.38) (1.38) POWER2 –0.005 –0.005 (–0.68) (–0.68) POWER3 0.014** 0.014** (2.16) (2.16) LEVERAGE –0.057*** –0.053*** –0.057*** –0.053*** (–2.88) (–2.66) (–2.88) (–2.66) SIZE 0.049*** 0.048*** 0.049*** 0.048*** (14.00) (13.64) (14.00) (13.64) ROA 0.826*** 0.824*** 0.831*** 0.825*** (19.06) (18.59) (19.00) (18.59) TOTAL_ASSET_ 0.007** 0.007* 0.007** 0.007* GROWTH (2.04) (1.83) (2.04) (1.83) INVENTORY_RATIO 0.052** 0.053** 0.052** 0.053** (2.10) (2.11) (2.10) (2.11) BIG4 0.039*** 0.036*** 0.039*** 0.036*** (2.79) (2.57) (2.79) (2.57) LARGEST_SHARE 0.001*** 0.001*** 0.001*** 0.001*** (3.00) (3.06) (3.00) (3.06) ConStant 5.317*** 5.315*** 5.326*** 5.324*** (71.82) (70.37) (70.44) (69.04) YEAR/IND Control Control Control Control a djusted R 0.481 0.481 0.482 0.483 F 57.68 49.97 57.68 49.97 N 1,107 1,107 1,107 1,107 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for two-tailed test. 4.3. Empirical results analysis for hypotheses H2 and H3 As Table 9 shows, in the growth stage the coefficients on POWER1 and POWER2 are signifi- cantly positive (0.007 with t = 1.79, 0.006 with t = 2.22), while the coefficient of POWER3 is significantly negative (–0.008 with t = –2.54). This verifies the hypotheses H2-1, H2-2 and H2-3. However, in the non-growth stage, the coefficients on POWER2 and POWER3 are sig- nificantly negative (–0.018 with t = –2.52, –0.016 with t = –2.44), which verifies H3-1 and H3-2. But POWER1 is not significantly correlated with internal control quality in the non-growth stage, which cannot verify H3-3. We conclude in the growth stage: when the CEO is fully granted structure and expert power, it is beneficial to exert a power guarantee function. In the non-growth stage, the higher POWER2 may result in higher agency costs, which causes the internal control system to become part of agency problems. The reason why H3-3 is not supported is that, in the non-growth stage, family CEO, professional manager and non-nuclear family member’s entrance into the firm lead to more complex CEO com- position and aggravate agency conflicts, we do not eliminate different CEO type effects. Thus, it may influence the result’s significance. The coefficients of SIZE, ROA and INVENTORY_ RATIO are significantly positive, which implies firms with larger size, better performance and higher inventory ratio have higher demand for internal control quality. These conclusions are consistent with the opinions of Gong, Ke, and Yu (2013). 226 M. HU AND S. GAN Table 9. regression results of Ceo power and internal control quality in different life cycle stages. Full samples Growth stage Non-growth stage Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) POWER1 0.008** 0.007* 0.007 (2.23) (1.79) (1.07) POWER2 0.004 0.006** –0.018** (1.10) (2.22) (–2.52) POWER3 –0.009*** –0.008*** –0.016** (–3.35) (–2.54) (–2.44) LARGEST_SHARE 0.007 0.009 0.003 (0.53) (0.65) (0.11) LEVERAGE –0.023* –0.013 –0.047* (–1.77) (–0.92) (–1.93) SIZE 0.036*** 0.033*** 0.038*** (11.27) (9.25) (6.74) ROA 0.843*** 0.959*** 0.647*** (10.06) (17.24) (4.51) AGE –0.001** –0.000 –0.003*** (–2.16) (–0.24) (–3.80) INVENTORY_RATIO 0.045*** 0.037* 0.076*** (2.65) (1.79) (2.75) REVENUE_GROWTH –0.000*** –0.000*** 0.001 (–12.96) (–13.76) (1.28) BIG4 0.037** 0.043** 0.014 (1.97) (2.14) (0.84) CONSTANT 5.651*** 5.682*** 5.622*** (83.35) (73.36) (45.55) YEAR/IND Control Control Control a djusted R 0.368 0.368 0.416 F 48.76 51.39 17.02 N 3,635 2,797 838 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. 4.4. CEO type moderation effect result analysis 4.4.1. Actual controller as CEO We verify the impact of actual controller as CEO with results shown in Table 10. Panel A presents that in the sample of actual controller as CEO, the mean values of POWER1, POWER2 and POWER3 are significantly higher than other types of CEO, but the mean value of internal control quality is significantly lower than other types of CEO. Panel B indicates the coefficient on SELF is significantly negative (–0.013 with t = –1.66), which illustrates the actual controller as CEO has a direct negative effect on the internal control quality. The result indicates actual controller as CEO is not helpful to improve the internal control quality. The reasons may be as follows. First, actual controller as CEO exacerbates interest conflicts between family owners and minority shareholders. To prevent other party’s supervision and intervention, family owners may prefer weaker internal control quality. Second, due to the special feature of family firms, actual controllers are more willing to adopt an informal governance mechanism such as kinship, trust or emotion to solve the agency problem. However, internal control, as a mean of formal governance mechanism, is just to meet the supervision need. Hence, fewer resources are devoted to establishing an internal control system. Furthermore, the coefficient of SELF multiplied by POWER2 is significantly positive (0.009 with t = 1.74), illustrating SELF has a positive moderating effect on the relationship of POWER2 and internal control quality. CHINA JOURNAL OF ACCOUNTING STUDIES 227 Table 10. regression result of actual controller as Ceo on Ceo power and internal control quality. Panel A: T test Self other t value N 1,478 2,157 ln(ICINDEX) 6.51 6.53 −3.17*** POWER1 0.73 0.37 21.43*** POWER2 0.56 0.41 16.41*** POWER3 1.43 0.95 24.42*** Panel B: Regression results of CEO type effect f ull samples growth stage non-growth stage Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) SELF –0.013* –0.009 –0.017 (–1.66) (–0.94) (–1.01) POWER1 0.007* 0.006 0.004 (1.80) (1.46) (0.55) POWER2 0.000 0.001 –0.021** (0.06) (1.21) (–2.33) POWER3 –0.014*** –0.011** –0.024*** (–2.97) (–2.44) (–2.65) SELF×POWER1 0.001 0.001 0.008 (0.24) (0.06) (0.57) SELF×POWER2 0.009* 0.008 0.006 (1.74) (1.54) (0.45) SELF×POWER3 0.007 0.004 0.012 (1.40) (0.75) (0.99) LEVERAGE –0.022** –0.013 –0.049** (–2.14) (–1.08) (–2.34) SIZE 0.036*** 0.033*** 0.041*** (17.38) (13.61) (9.48) ROA 0.845*** 0.961*** 0.648*** (30.74) (27.63) (13.67) LARGEST_SHARE 0.008 0.000 0.000 (0.73) (0.86) (0.15) BIG4 0.037*** 0.044*** 0.013 (3.61) (3.79) (0.58) INVENTORY_RATIO 0.047*** 0.038** 0.077*** (3.30) (2.32) (2.78) REVENUE_GROWTH –0.000*** –0.000*** 0.001 (–2.67) (–2.97) (0.65) AGE –0.001*** 0.000 –0.003*** (–2.76) (0.05) (–3.88) CONSTANT 5.659*** 5.689*** 5.628*** (129.92) (109.51) (63.50) YEAR/IND Control Control Control N 3,635 2,797 838 F 84.49 64.84 23.26 a djust R 0.365 0.363 0.399 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. 4.4.2. Kinship network of family CEO To clarify family CEO’s impact, FCEOTYPE takes 4, 3, 2, 1 successively based on the kinship network. Table 11 shows, in the full sample, kinship network is negative with internal control quality (–0.023 with t = –1.81), which illustrates the closer kinship network of family CEO with actual controller may lead to lower internal control quality. A closer kinship network of family CEO with actual controller may intensify agency coni fl cts between family owners and minority shareholders. Furthermore, the coefficient of FCEOTYPE multiplied by POWER3 is significantly positive (0.023 with t = 2.15), which indicates that the closer kinship network 228 M. HU AND S. GAN Table 11. f amily Ceo ’s kinship network effect on Ceo power and internal control quality. Full samples Growth stage Non-growth stage Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) FCEOTYPE –0.023* –0.015 –0.048 (–1.81) (–1.09) (–1.09) POWER1 0.029 0.047 –0.008 (0.96) (1.28) (–0.09) POWER2 –0.015 –0.005 –0.118 (–0.40) (–0.10) (–0.67) POWER3 –0.063* –0.059 –0.094 (–1.74) (–1.49) (–0.63) FCEOTYPE×POWER1 –0.009 –0.018 0.001 (–1.10) (–1.59) (0.04) FCEOTYPE×POWER2 0.007 0.005 0.047 (0.68) (0.39) (1.13) FCEOTYPE×POWER3 0.023** 0.019 0.046 (2.15) (1.62) (1.16) LEVERAGE 0.021 0.034 0.029 (0.58) (0.77) (0.37) SIZE 0.033*** 0.034*** 0.019 (3.94) (3.38) (0.88) ROA 1.135*** 1.134*** 1.141*** (9.96) (8.25) (4.06) REVENUE_GROWTH 0.001 0.001 –0.002 (0.42) (0.45) (–0.78) LARGEST_SHARE 0.076 0.121** 0.037 (1.61) (2.22) (0.28) INVENTORY_RATIO –0.001 –0.034 0.035 (–0.42) (–0.69) (0.27) BIG4 –0.028 –0.011 –0.082 (–0.89) (–0.27) (–1.09) AGE –0.001 –0.001 –0.003 (–0.72) (–0.56) (–0.71) CONSTANT 5.718*** 5.662*** 6.178*** (30.07) (25.62) (10.78) YEAR/IND Control Control Control N 275 210 65 F 7.57 6.15 2.08 a djust R 0.376 0.381 0.301 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. can increase the CEO ownership power’s negative effect on internal control quality. It indi- cates that the entrenchment eec ff t dominates the concern of a family member or the owner for the firm’s long-term welfare. 4.5. Robustness tests To ensure our results are consistent across different sample and model specifications, we conduct several robustness tests. First, we define the CEO power as a comprehensive index by mean value, which takes the value 1 if it is higher than the industry mean value, otherwise 0, and rerun the model in Equations (2). Next, we define the internal control index higher than industry mean value as 1, otherwise 0, and the CEO power as a comprehensive index by the principal component analysis method. Non-tabulated results are basically consistent with the primary tests. CHINA JOURNAL OF ACCOUNTING STUDIES 229 Table 12. robustness checks using two-stage least square regression procedures. Panel A: CEO power’s lag value as instrument variable f ull samples growth stage non-growth stage Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) POWER1 0.009* 0.011* –0.008 (1.94) (1.86) (–0.85) POWER2 –0.027 –0.004 –0.059** (–1.61) (–0.18) (–2.21) POWER3 –0.002 –0.005 0.004 (–0.61) (–0.90) (0.36) LARGEST_SHARE 0.014 0.009 0.015 (1.13) (0.65) (0.62) LEVERAGE 0.011* 0.022 –0.026 (0.90) (1.51) (–1.08) SIZE 0.031*** 0.026*** 0.036*** (12.21) (8.28) (7.46) ROA 1.013*** 1.101*** 0.856*** (28.27) (25.16) (13.12) AGE –0.002*** –0.000 –0.003** (–2.87) (–0.31) (–2.16) INVENTORY_RATIO 0.036** 0.022 0.065** (2.15) (1.10) (2.16) REVENUE_GROWTH 0.000 0.000 0.000 (0.45) (0.09) (0.60) BIG4 0.047*** 0.063*** 0.015 (3.69) (4.15) (0.68) CONSTANT 5.791*** 5.851*** 5.751*** (107.12) (83.79) (59.85) YEAR/IND Control Control Control a djust R 0.401 0.431 0.392 Panel B: CEO power’s median value as instrument variable f ull samples growth stage non-growth stage Coefficient Coefficient Coefficient Variables (t-value) (t-value) (t-value) POWER1 0.008* 0.008* 0.002 (1.80) (1.65) (0.29) POWER2 0.003 0.012* –0.022* (0.51) (1.69) (–1.73) POWER3 –0.000 –0.002 –0.002 (–0.01) (–0.28) (–0.30) LARGEST_SHARE 0.011 0.011 0.009 (0.92) (0.92) (0.42) LEVERAGE –0.013 –0.003 –0.047** (–1.24) (–0.22) (–2.18) SIZE 0.036*** 0.034*** 0.031*** (16.84) (13.21) (9.36) ROA 0.825*** 0.968*** 0.601*** (28.45) (26.32) (11.71) AGE –0.001*** –0.001 –0.003*** (–2.89) (–0.83) (–3.01) INVENTORY_RATIO 0.041*** 0.035** 0.065** (2.67) (2.00) (2.29) REVENUE_GROWTH 0.000 0.000 0.000 (0.40) (0.07) (0.50) CONSTANT 5.64*** 5.67*** 5.61*** (122.94) (102.38) (62.74) YEAR/IND Control Control Control a djust R 0.341 0.342 0.381 (Continued) 230 M. HU AND S. GAN Table 12. (Continued). Panel C: Controlling company level fixed effect result f ull samples growth stage non-growth stage Coefficient Coefficient Coefficient (t-value) (t-value) (t-value) Variables POWER1 0.015*** 0.011* 0.035*** (2.91) (1.73) (2.62) POWER2 –0.003 0.003 –0.021* (–0.62) (0.53) (–1.66) POWER3 –0.012*** –0.003 –0.041*** (–2.70) (–0.52) (–3.50) LARGEST_SHARE –0.055 –0.087 0.044 (–1.12) (–1.25) (0.41) LEVERAGE –0.009 0.023 0.013 (–0.39) (0.76) (0.18) SIZE 0.044*** 0.045*** 0.021 (5.671) (4.231) (0.851) ROA 0.791*** 0.931*** 0.591*** (19.82) (16.27) (8.19) AGE 0.021* 0.008 0.015 (1.82) (0.58) (0.64) INVENTORY_RATIO –0.069** –0.065 –0.055 (–2.20) (–1.58) (–0.91) REVENUE_GROWTH –0.000 –0.001 –0.000 (–0.29) (–0.31) (–0.28) BIG4 –0.006 –0.035 0.073 (–0.22) (–0.98) (0.79) CONSTANT 5.431*** 5.451*** 5.911*** (31.46) (22.80) (10.98) YEAR/IND Control Control Control a djust R 0.478 0.474 0.478 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. 5. Endogeneity issues between CEO power and internal control quality 5.1. The potential causality between CEO power and internal control quality There appears to be a causal relationship between CEO power and internal control quality, because lower internal control quality may lead to higher CEO power. In this section, we address the potential endogeneity problems in two different ways. First, we use instrumental variable methods to isolate the effects of CEO power on internal control quality. We use CEO power’s lag and median value, respectively, as instrument variables, and then we do a two- stage least square regression in Equation (2). The empirical results are illustrated in Table 12. Panel A and Panel B indicate that the CEO power has a direct effect on internal control quality. Second, we control company level fixed effects. Panel C shows the coefficient of POWER1 is significantly positive (0.011 with t = 1.73) in the growth stage, POWER2 is significantly neg- ative (–0.021 with t = –1.66) and POWER3 is significantly negative (–0.041 with t = –3.50) in the non-growth stage, suggesting that listed family firms in China are likely to have a higher level of internal control quality relative to the structure power in the growth stage and lower internal control quality related to the expert power and ownership power in the non-growth stage. We conclude that CEO power has a direct effect on internal control quality. 5.2. Sample self-selection bias issues In addition, because we choose China-listed family firms as a sample, there may be issues of sample selection bias. We test this using the propensity score matching method. The CHINA JOURNAL OF ACCOUNTING STUDIES 231 Table 13. PSM results. Growth stage Non-growth stage Variables Coefficient Z value Coefficient Z value POWER1 0.004 0.04 –0.053 –0.23 POWER2 0.121 0.79 –0.591 –2.08** POWER3 –0.241 –2.04** –0.222 –0.94 LEVERAGE –0.191 –0.43 –0.893 –1.29 SIZE 0.401 4.17*** 0.592 3.86*** ROA 23.801 12.68*** 24.174 7.60*** TOTAL_ASSET_GROWTH 0.122 0.59 0.091 0.39 INVENTORY_RATIO 0.543 0.93 1.492 1.65* REVENUE_GROWTH –0.014 –0.50 0.021 0.90 BIG4 0.691 1.19 –0.243 –0.21 LARGEST_SHARE –0.004 –0.85 –0.008 –0.84 AGE 0.002 0.10 –0.039 –1.45 CONSTANT –8.051 –3.99*** –11.231 –3.57*** YEAR/IND Control Control Pseudo R 0.221 0.362 note: ***, ** and * represent the 1%, 5% and 10% levels of significance respectively for the two-tailed test. empirical results are given in Table 13. The coefficient of POWER3 is significantly negative (–0.241 with z = –2.04) in the growth stage, and the coefficient of POWER2 is significantly negative (–0.591 with z = –2.08) in the non-growth stage. These results are similar to the primary tests. These results illustrate that the conclusion is robust. 6. Conclusions Our findings have some practical implications. Because of the power heterogeneity resulting from the complex composition of CEOs, actual controllers or their family members as CEO (called ‘family controlled and family managed’) can alleviate the conflicts of owner and man - ager, but aggravate the conflicts of interest between family owners and minority sharehold - ers. It indicates that the entrenchment effect dominates a family owner’s concern for the long-term welfare of firms, which is consistent with actual controllers or their family members as CEO preferring lower internal control quality in order to entrench themselves. For family controlled and family managed firms, kinship, trust or emotion may replace the internal control governance mechanism to some extent. Therefore, for this type of family firm, internal control is present only to meet the needs of regulation and fewer resources are devoted to the implementation of internal control. The results show that the life cycle stages of listed family firms in China need to be identified and the CEO power source needs to be adjusted in the process of constructing internal control. Our study is subject to the following limitations. If the corporationh actual controllers or their family members as CEO preferring lower internal control quality in order to entrench themselvesbe more significant. And in the construction of life cycle stage index, breaking only the financial indicator and establishing a more reasonable classification standard may be more beneficial to the subsequent improvement. Despite these limitations, this study increases our understanding of the internal control quality of listed family firms in China. Acknowledgements We appreciate constructive comments from Professor Cong Wang and the CJAS editorial department, and participants at our presentations at Nanjing University. 232 M. HU AND S. GAN Disclosure statement No potential conflict of interest was reported by the authors. References Adizes, I. (1989). How and why corporation grow and die and what to do about it: Corporate life cycle. Englewood Cliffs, NJ: Prentice Hall. Aharony, J., Falk, H., & Yehuda, N. (2004). Corporate life cycle and the value relevance of cash flow versus accrual financial information. Tel Aviv University, Working Paper. Ali, A., Chen, T.Y., & Radhakrishnan, S. (2007). Corporate disclosures by family firms. Journal of Accounting and Economics, 44(1), 238–286. Allen, M.P. (1981). Managerial power and tenure in the large corporation. Social Forces, 60, 482–494. Anderson, R.C., & Reeb, D.M. (2003a). Founding-family ownership and firm performance: Evidence from the S&P 500. The Journal of Finance, 58, 1301–1328. Anderson, R.C., & Reeb, D.M. (2003b). Founding-family ownership, corporate diversification and firm leverage. Journal of Law and Economics, 46, 653–684. Bardhan, I., Lin, S., & Wu, S.L. (2015). The quality of internal control over financial reporting in family firms. Accounting Horizons, 29(1), 41–60. Black, E.L. (2000). Life-cycle stage effects on the incremental information content of earnings and cash flows. Brigham Young University, Working Paper. Bunkanwanicha, P., Fan, J.P.H., & Wiwattanakantang, Y. (2013). The value of marriage to family firms. Journal of Financial and Quantitative Analysis, 48, 611–636. Bushman, R.M., & Smith, A.J. (2001). Financial accounting information and corporate governance. Journal of Accounting and Economics, 32, 237–333. Chrisman, J.J., Chua, J.H., Kellermanns, F.W., & Chang, E.P.C. (2007). Are family managers agents or stewards: An exploratory study in privately held family firms. Journal of Business Research, 60, 1030– Davis, J.H., Schoorman, F.D., & Donaldson, L. (1997). Toward a stewardship theory of management. The Academy of Management Review, 22(1), 20–47. Dickinson, V. (2011). Cash flow patterns as a proxy for firm life cycle. The Accounting Review, 86, 1969– Doyle, J.T., Ge, W., & McVay, S. (2007). Accruals quality and internal control over financial reporting. The Accounting Review, 82, 1141–1170. Drazin, R., & Kazanjian, R.K. (1990). A reanalysis of Miller and Friesen`s life cycle data. Strategic Management Journal, 11, 319–325. Eisenhardt, K.M. (1989). Agency theory: An assessment and review. The Academy of Management Review, 14(1), 57–74. Eisenhardt, K.M., & Zbaracki, M.J. (1992). Strategic decision making. Strategic Management Journal, 13, 17–37. Fan, J.P.H., & Wong, T.J. (2002). Corporate ownership structure and the informativeness of accounting earnings in East Asia. Journal of Accounting and Economics, 33, 401–425. Fei, X. (1947). Rural China. Shanghai: Sanlian Press (In Chinese). Feng, M., Li, C., & McVay, S. (2009). Internal control and management guidance. Journal of Accounting and Economics, 48, 190–209. Finkelstein, S. (1992). Power in top management teams: Dimensions, measurement and validation. The Academy of Management Journal, 35, 505–538. Forbes China Rich List. (2014). Retrieved from https://www.forbes.com/sites/ceibs/2014/12/03/chinese- family-firms-the-good-the-bad-and-the-ugly/#6f9ce6456f9c Gomez-Mejia, L.R., & Wiseman, R.M. (2007). Does agency theory have universal relevance? A reply to Lubatkin, Lane, Collin, and Very. Journal of Organizational Behavior, 28(1), 81–88. Gong, G., Ke, B., & Yu, Y. (2013). Home country investor protection, ownership structure and cross- listed firms’ compliance with SOX-mandated internal control deficiency disclosures. Contemporary Accounting Research, 30, 1490–1523. CHINA JOURNAL OF ACCOUNTING STUDIES 233 Hambrick, D.C., & Mason, P.A. (1984). Upper echelons: The organization as a reflection of its top managers. Academy of Management Review, 9, 193–206. He, X., Li, J., & Chen, L. (2010). Family member composition and corporate governance efficiency: Empirical research from family firms. Nankai Business Review, 13, 149–160 (In Chinese). He, X., & Lian, Y. (2009). Family authority and firm value: Empirical evidence from family firms. Economic Research Journal, 4, 90–102 (In Chinese). Hu, M., Gan, S., & Lu, Y. (2015). Property rights, managerial power and internal control. Journal of Chongqing University, 21, 67–80 (In Chinese). Karra, N., Tracey, P., & Phillips, N. (2006). Altruism and agency in the family firm: Exploring the role of family, kinship, and ethnicity. Entrepreneurship Theory and Practice, 30, 861–877. Leone, A.J., & Liu, M. (2010). Accounting irregularities and executive turnover in founder-managed firms. The Accounting Review, 85(1), 287–314. Li, Y., Li, Z., & Tang, S. (2011). Life cycle stage, corporate governance and capital allocation efficiency. Nankai Business Review, 14, 110–121 (In Chinese). Lubatkin, M.H., Ling, Y., & Schulze, W.S. (2007). An organizational justice-based view of self-control and agency costs in family firms. Journal of Management Studies, 44, 955–971. Ma, Y., Li, H., & He, F. (2008). An analysis of high-tech enterprise corporate governance mechanism based on power game: From the corporate governance perspective. Reform of Economic System, 5, 58–61 (In Chinese). Mehrotra, V., Morck, R., Shim, J., & Wiwattanakantang, Y. (2013). Adoptive expectations: Rising sons in Japanese family firms. Journal of Financial Economics, 108, 840–854. Mintzberg, H., & Waters, J.A. (1990). Tracking strategy in an entrepreneurial firm. Family Business Review, 3, 465–499. Mullins, W., & Schoar, A. (2016). How do CEOs see their role? Management philosophies and styles in family and non-family firms. Journal of Financial Economics, 119(1), 24–43. Mustakallio, M., Autio, E., & Zahra, S.A. (2002). Relational and contractual governance in family firms: Effects on strategic decision making. Family Business Review, 15, 205–222. Schulze, W.S., Lubatkin, M.H., Dino, R.N., & Buchholtz, A.K. (2001). Agency relationships in family firms: Theory and evidence. Organization Science, 12, 99–110. Skaife, H.A., Collins, D.W., Kinney, W.R., & LaFond, R. (2007). The ee ff ct of SOX internal control dec fi iencies and their remediation on accrual quality. The Accounting Review, 83(1), 217–250. Skaife, H.A., Veenman, D., & Wangerin, D. (2013). Internal control over financial reporting and managerial rent extraction: Evidence from the profitability of insider trading. Journal of Accounting and Economics, 55(1), 91–110. Smith, C.W., & Warner, J.B. (1979). On financial contracting: An analysis of bond covenants. Journal of Financial Economics, 7, 117–161. Su, Q. (2007). Family firm’s manager behavior choice based on the agency and stewardship theory. Foreign Economics and Management, 29, 51–58 (In Chinese). Van den Berghe, L.A.A., & Carchon, S. (2003). Agency relations within the family business system: An exploratory approach. Corporate Governance: An International Review, 11, 171–180. Wang, D. (2006). Founding family ownership and earnings quality. Journal of Accounting Research, 44, 619–656. Wang, M., Xu, M., & Wang, H. (2014). Does altruistic behavior reduce the agency cost? An empirical study based on kin altruistic behavior in family business. Economic Research Journal, 3, 144–157 (In Chinese). Xie, Z. (2009). The nature and structure of internal control. Accounting Research, 12, 70–75 (In Chinese). Xu, J., & Lv, C. (2011). The nature of the top managers in clannish firms and the quality of earnings therein. Management World, 1, 112–120 (In Chinese). Zahra, S.A., & Pearce, J.A. (1989). Boards of directors and corporate financial performance: A review and integrative model. Journal of Management, 15, 291–334. Zhang, Y., & Zheng, H. (2010). An investigation and analysis of the effectiveness of internal control and its influencing factors in China. Audit Research, 1, 75–81 (In Chinese). Zhao, Y., & Lv, C. (2015). Family or money? The impact of family relationship on compensation contract in family firms. Accounting Research Journal, 8, 32–40 (In Chinese).
China Journal of Accounting Studies – Taylor & Francis
Published: Apr 3, 2017
Keywords: CEO power; internal control quality; kinship; life cycle stage
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.