Abstract
China Journal of Accounting Studies, 2013 Vol. 1, Nos. 3–4, 236–259, http://dx.doi.org/10.1080/21697221.2013.867401 Why do state-owned enterprises over-invest? Government intervention or managerial entrenchment a b Jun Bai * and Lishuai Lian a b School of Economics and Management, Shihezi University, Xinjiang 832000, China; School of Management, Fudan University, Shanghai, 200433, China In a transition economy, corporate investment decisions are affected not only by managerial discretion, but also by government intervention. Using the data of pub- licly listed state-owned enterprises (SOEs) in China, we investigate how government intervention and corporate managerial entrenchment affect over-investment. The results show that both the policy burden from government intervention and rent- seeking due to managerial entrenchment can lead to over-investments, and these two effects appear to be complementary to each other. With a weak government inter- vention, managerial discretion is greater and management behavior tends toward opportunism. Keywords: government intervention; managerial entrenchment; over-investment 1. Introduction China’s economy has been growing rapidly since the start of its economic reform in 1978. Its GDP reached US$6.01 trillion in 2010 (National Bureau of Statistics of China, 2011) and it is now the world’s second largest economy. The China’s growth pattern can be described by the investment-growth model (Zhang, 2003). The invest- ment-to-GDP ratio in China, with continuous rapid growth from 36% to 48% in the period of 2001 and 2010, is high relative to the world average (23.7% in 2010), and it is also higher than most other East Asian countries, which are also known for relying on capital accumulation for their growth (27% in 2010). Despite this, a high invest- ment-to-GDP ratio does not necessarily correspond to optimal investment as firms may over-invest, e.g. firms’ engagement in redundant constructions (Tang, Zhou, & Ma, 2007, 2010). This problem is particularly severe among state-owned enterprises (SOEs), where investment efficiency is significantly lower than that of domestic private or foreign-owned firms (Dollar & Wei, 2007). Traditional investment theory suggests that over-investment is rooted in efficiency loss for an internal agency chain. In modern firms, where ownership and control are separate, managers, acting as agents, hold the significant control rights over firms’ asset allocation decisions (Shleifer & Vishny, 1997). Managers may pursue their own private goals that might be in conflict with those of outside shareholders. Managers can take wasteful, negative net present value investment projects as a result of their desire to derive personal benefits from controlling more assets, which leads to over-investment (e.g., Jensen, 1986, 1993; Stulz, 1990). *Corresponding author. Email: baijun@shzu.edu.cn Paper accepted by Tong Yu. © 2013 Accounting Society of China China Journal of Accounting Studies 237 In a transition economy, the government plays a vital role in the business activities of firms, and more profoundly impacts SOEs’ investment decisions (Chen, Sun, Tang, &Wu, 2011). The enhanced government control over SOEs, due to fiscal decentraliza- tion reform (Qian & Weingast, 1997) and the reform of political promotion mechanisms (Blanchard & Shleifer, 2001; Li & Zhou, 2005), allows the government to impose eco- nomic growth and social stability targets on firms by intervening in their investment decisions. This makes the investment decisions deviate from maximizing the benefits of shareholders, thereby promoting SOEs’ over-investment. Furthermore, during gradual decentralization in the SOE reformation, the role of corporate managers becomes increasingly important. Managers may maximize their rent-seeking through investment expansion, which yields SOEs’ over-investment (Wei & Liu, 2007; Xin, Lin, & Wang, 2007; Zhang, Wu, & Wang, 2010). Government intervention and managerial entrench- ment have become the primary forces driving SOEs’ over-investment, and these two forces potentially interact with each other. Therefore, to understand the influences of government intervention and managerial entrenchment on over-investment, the paper examines the following. (1) What are the impacts of these two forces on SOEs’ investment decisions? (2) How do they interact with each other during the process? Differently put, are they supplementary or complementary to each other? (3) Which one is the dominant force? Is there a difference in managerial entrenchment under heterogeneous levels of government interventions? To address the above questions, the paper investigates the impacts of policy burden from government intervention and rent-seeking behavior that originated from manage- rial entrenchment on firms’ over-investment. We find that in SOEs, both government intervention and managerial entrenchment can yield over-investment, but the two have a complementary relationship in affecting over-investment. More powerful government interventions suppress managerial entrenchment, and management behavior tends toward opportunism in a weak government intervention environment. The paper explores the interactions between the government and firm management from an investment perspective and contributes to the literature in three ways. First, we provide microscopic evidence on how government behavior affects economic growth, which enables us to better understand the mechanism underlying the impacts of govern- ment behavior on microscopic financial decisions, thereby enriching the extant literature on the relationship between government and firms. China’s economy is in a transition stage, where various government policies are undergoing reforms. The relationship between government behavior and economic growth has been the focus of academic research, but the micro-channels and mechanisms through which government affects economic growth have not been well-understood. Herein, by incorporating government behavior into the firm micro-financial decision-making framework, this paper investi- gates the impacts of mutual influences between government intervention and manage- ment behavior on firms’ investment decisions, providing a new theoretical perspective to better understand current government behavior in China and its economic consequence. Second, by measuring the ‘policy burden’ on firms, this paper evaluates govern- ment behavior from a microscopic perspective, bridging studies in macro-public and micro-corporate governance. Following La Porta, Lopez-de-Silanes, Shleifer, and Vish- ny, (1998), studies have recognized the importance of the impacts of government behavior and institutional environment on firm’s financial decisions. However, little has been done to quantify government intervention. Based on the ratio and structure of government ownership and focusing on the Chinese financial market, researchers have attempted to quantify government behavior mainly from two aspects. One is by 238 Bai and Lian assessing government equity (Sun & Tong, 2003; Tang et al., 2010;Wu, 2009;Xia & Chen, 2007). The other is using pyramid structure as the substitution variable of government decentralization (Fan, Wong, & Zhang, 2012) or intervention (Cheng, Xia, &Yu, 2008; Zhong, Ran, & Wen, 2010). In line with the view that the policy burden reflects the direct economic consequence of government intervention in SOEs (Lin & Li, 2008), we employ policy burden as a proxy for government intervention, which is a new way to quantify government behavior at the micro level. Finally, our findings shed new light on the literature that examines distortional investment behavior due to the traditional agency conflicts between shareholders and managers as well as the conflicts between shareholders and debt holders in mature mar- kets. The traditional investment theory suggests that corporate investment is indepen- dent of the government regulatory environment. Studies focus on the impacts of the internal stockholder, manager and external investor agency relationships on firm finan- cial decision-making (Jensen & Meckling, 1976; Shleifer & Vishny, 1986, 1989). We show that governments in transition economies like China can expropriate the firms by intervening in their investment decisions. Meanwhile government activities also restrict the managerial entrenchment. By considering the government intervention and manage- rial entrenchment in the transitional economy of China, the findings facilitate our understanding of these institutional features’ impacts on SOEs’ investment activities. The remainder of the paper proceeds as follows. Section 2 discusses the institutional background and the theoretical framework of the study. Section 3 presents the research design, introduces the samples, data and models adopted in the paper. Section 4 reports the empirical findings and Section 5 concludes. 2. Institutional background and theoretical analysis 2.1. Inducements of over-investment Investment decisions are not only vital for the survival and growth of firms at the micro level, but also they serve an economic engine at the macro-economic level. In a perfect capital market, a firm’s investment policy is solely dependent on its investment oppor- tunities (Fazzari, Hubbard, & Petersen, 1988; Modigliani & Miller, 1958). However, in reality, firms would deviate from the optimal investment level for various reasons. Investment decisions reflect not only the agency relationship within the firms (Bushman & Smith, 2001; Stein, 2003) but also government behavior and the relationship between government and firms (Macroeconomic Study Group, CCER, Peking Univer- sity, 2004). At the firm level, agency problems would lead managers to over-invest when they pursue their own interests (Jensen, 1986). In the government and firm rela- tionship, to accomplish social and political goals, the government has strong incentives to intervene in firms’ decision-making, which may also lead to over-investment. The problem is particularly severe for SOEs (Chen, Sun, Tang, & Wu, 2011). 2.1.1. Managerial entrenchment and over-investment At the firm level, the agency problem stems from the separation of ownership and control. When managers end up with significant control rights over how to allocate investors’ funds, they may use their discretion for rent-seeking. This incurs great costs to the firms, and thereby reduces shareholders’ wealth. The agency theory suggests that managers tend to expand investments and are motivated to invest free cash in non-profit projects for their personal benefits. In other words, managers have incentives China Journal of Accounting Studies 239 to expand investment beyond the optimal level. To interpret over-investment caused by the agency problem, researchers have proposed various theoretical frameworks, such as the free cash-flow agency (Jensen, 1986), empire-building (Stulz, 1990), entrenchment (Shleifer & Vishny, 1989) and overconfidence (Roll, 1986) hypotheses. Generally, three principal explanations have been presented for why managers over-invest. First, Jensen (1986) argues that managers have incentives to over-invest because payouts of free cash to shareholders reduce the resources under their control, thereby reducing managers’ power, and making it more likely that the shareholders will incur the monitoring of the capital markets when a firm needs to raise fund. Second, by making manager-specific investment that often beyond its optimal size, managers can reduce the probability of being replaced (Shleifer & Vishny, 1989). Third, Stulz (1990) shows that over-confident managers tend to invest to the maximum extent when managers have discretion over firms’ decision-making. 2.1.2. Government intervention and over-investment One cannot ignore the effect of government behavior on corporate investment decisions. Government may internalize non-economic efficiency goals into firms’ investment activi- ties by intervening in management decisions, with both effects of ‘grabbing hands’ and ‘helping hands’. For the former, government intervenes in firms’ investment activities to increase fiscal revenue, to improve social welfare, to maintain social stability and to achieve other social and political goals, which yields deviations in investment decisions from market efficiency and encroaches on and reduces shareholders’ wealth. The latter manifests itself such that government is intimately involved in promoting economic activity, and supports firms with close ties to it (Frye & Shleifer, 1997). Meanwhile, in a market economy, a firm may be short-sighted and this can be corrected by government intervention. In particular, when an internal management agency problem is serious, government control and supervision can inhibit managerial entrenchment, thereby protecting external investors’ interests from being abused. When a government takes actions to improve social welfare by expropriating firms, and firm managers also expropriate firms to maximizing their own benefits, government agency and management agency problems will occur simultaneously (Stulz, 2005). These two agency problems are neither irrelevant nor independent, but intertwined. Government behavior affects or restricts the personal benefits of managers, while managers are motivated to prevent such predatory government behavior to protect their personal interests. Therefore, firm investment decisions are a balanced outcome from interactions between government and management. 2.2. Why do China’s SOEs over-invest? Theoretical analysis 2.2.1. Economic reform and government intervention in China China has experienced the devolution of fiscal authority from central to local governments since 1978. The major objectives of fiscal reform were to make localities fiscally self-sufficient, to reduce the central state’s own fiscal burden including subsidizing inefficient firms, to maintain the social stability, and to provide incentives for local authorities to promote economic development. During the reform, localities became independent in terms of fiscal entities that had both responsibility for local expenditures and the unprecedented right to use the revenue that they retained (Oi, 1992). Local governments gained powers such as financial discretion and management 240 Bai and Lian of social public economy and also assumed social goals such as economic development, employment, social pension and social stability. To adapt to economic reform, in the early 1980s, China introduced the government official incentive system, where the Chinese central government was ready to reward or punish local officials on the basis of their economic performance in order to motivate them to promote the local economy (Blanchard & Shleifer, 2001; Li & Zhou, 2005). The central government has the power to appoint and dismiss local officials, and has exercised such power both to support the local officials whose regions have performed well economically and to discipline local officials who have followed anti-growth policies. In this context, local government officials are motivated to promote the local economy by generating fiscal revenue, GDP growth and social stability, among other goals, to gain political advance- ment opportunities. Firms’ investment not only boosts local economic growth but also increases employment opportunities, maintains social stability and facilitates govern- ment (officials’) performance evaluation goals, which facilitate political objectives. Therefore, investment may be the best option for local governments under competitive pressure to enhance their fiscal share and performance (Macroeconomic Study Group, CCER, Peking University, 2004). Meanwhile, inherent in gradual reforms, the govern- ment maintains ‘super control’ over SOEs. In order to promote economic growth and maintain social stability, the government (officials) has strong incentives to intervene in firms under the jurisdiction, especially SOEs, leading to the distortions of investment such as over-investment (Chen, Sun, Tang, & Wu, 2011), and SOEs bear the policy burden from multiple government objectives. Government intervention affects firms’ investment decisions in two different ways. One is referred to as a ‘protective effect’ on companies in a particular jurisdiction. On one hand, the bureaucrats’ political promotions rely on the performance of the SOEs under their jurisdiction. Therefore, they have responsibilities to support SOEs that undertake more policy burdens through providing resources and investment opportuni- ties, and subsidizing the SOEs when they are in trouble. In addition, when the legal systems are not well developed, the government can provide guarantees to the SOEs’ partners by using its own creditworthiness for implementing firm contracts, and help the SOEs be in a favorable position in the market competition. On the other hand, by monitoring managers and providing incentive plans for them, the government can effi- ciently mitigate the agency problems between the managers and outside investors. When the formal mechanisms, such as the legal mechanisms, are not efficient enough for the outside investors to protect themselves from expropriation by managers, the monitoring role played by the government can be viewed as an informal alternative (Zhong et al., 2010). Second, high government intervention in SOEs may distort their investment decisions when government (officials) has incentives to expropriate resources for its own goals, which yields a government-to-business ‘predatory effect’. The government (officials) may impose its own goals such as economic growth, social stability due to either fiscal reform or the previously mentioned political promotion, on SOEs, which leads these firms to over-invest simply because over-investment can create more jobs and promote GDP growth (Tang et al., 2010; Zhang & Wang, 2010). 2.2.2. SOE reformation and managerial entrenchment During the economic reforms in the 1980s, the Chinese government launched a program that decentralized the managerial decision rights of SOEs from the central government down to the local firm level. The owner of an SOE, a governmental China Journal of Accounting Studies 241 bureaucrat, typically faces decision-making constraints due to insufficient expertise and information, and thus allocates some decision rights to SOE managers. The SOE refor- mation went through four stages. The first stage ran from 1979 to 1983 with the major goals of administrative decentralization and profit retention, and in the second stage, SOEs were required to pay taxes instead of turning in profits, while the funding for SOEs’ capital investment, instead of being allocated directly from government financial reserves, had to come through bank loans (Sun & Tong, 2003). One of the major goals of the first two stages was to make the SOE mangers take responsibilities for their losses, and in this process, the SOE managers began to obtain some control rights from the government. In the third stage (1987–1992) with a Contractual Management Sys- tem, the government gave managers a free hand to run their operations, that is, the sep- aration of government ownership from control of SOEs’ operations (Sun & Tong, 2003). In the fourth stage of corporatization, the government granted significant con- trolling rights to SOE managers, except for the decision rights concerning M&A and the disposal of shares and assets, as well as the decision rights on the appointment of CEOs (Fan, Wong, & Zhang, 2007). However, empowered managers can expropriate substantial gains from the SOEs, resulting in severe agency costs. This is because, unlike a private firm, an SOE does not have a ‘true’ owner looking after the firm’s interests (Fan et al., 2012). Although the State-owned Assets Supervision and Adminis- tration Commission (SASAC) of the State Council and local governments were, since 2003, to take responsibility for managing the remaining SOEs, including appointing top executives and approving any mergers or sales of stock or assets and so on, it cannot mitigate the problem of ‘the absence of ownership’ in SOEs completely, and it is even possible that the officials of SASAC could make some concession to the managers in SOEs. Therefore, when empowered managers try to pursue their own benefits through investment decisions, over-investment would be their rational option (Wei & Liu, 2007; Xin et al., 2007; Zhang et al., 2010). When government intervenes in corporate investment decisions, managers’ responses to government actions potentially result in two types of relationship. One is a ‘complementary relationship’ between management and government. Predatory effects from government actions not only harm outside investor interests but also encroach on managers’ vested interests. Thus, to protect their own interests, managers are motivated to prevent government intervention, which yields a reciprocal substitutive relationship between the two entities. The other is a ‘supplementary relationship’ between manage- ment and government. From the above, under institutional fiscal decentralization, the government (officials) is interdependent with the firms under its jurisdiction. To get resources and investment opportunities from the government, managers may intention- ally meet government (officials’) economic development needs through investments; thus, SOEs become indispensable to local economic development and employment stability (Chen & Li, 2012). Therefore, the firms’ investment behavior results from both managerial entrenchment and government intervention. 2.2.3. Interaction between government intervention and managerial entrenchment Under the economic transition, fiscal decentralization and gradual reform of SOEs yielded that local governments were deeply involved in local economic development. Furthermore, local protectionism and regional market segmentation exists commonly, leading to significant variances in government impacts on SOEs. Within heterogeneous levels of the government intervention environment, government behavior and its 242 Bai and Lian impacts on managerial entrenchment evolve in different ways. In an environment where there is more severe government intervention, there will be less discretion for managers to make investment decisions due to government constraints and supervision, and thereby limiting their over-investment motivation. Hence, there are two options for corporate managers. First, they take actions to decrease expropriation by the govern- ment. Second, managers may cater to the needs of the government. Owing to the inter- connected and interdependent relationships between government (officials) and SOEs under its jurisdiction, managers of SOEs have to cater to the needs of the government (officials), so as to acquire the resources and investment opportunities provided by the government. In an environment with a relatively low government intervention, reduced government intervention is associated with an increased discretion in decision-making for the management; consequently, managers’ capacity to engage within government also increases. In this case, managers tend to behave in two different ways. One is opportunistic. Owing to the weaker government intervention and managerial entrench- ment when internal governance fails, while an effective market discipline mechanism is unavailable, unfettered managers tend to over-invest. The other behavioral tendency is market efficiency. If a firm is in a favorable market environment, management behavior is restricted by market forces, and the manager will follow market efficiency standards for investment decision-making. In summary, the economic transition in China yields government ‘super control’ over SOEs, and the government intervenes in SOEs’ investment decision-making to promote its political objectives. Such control over investment activities leads to devia- tions from the objective to maximize shareholders’ wealth, resulting in over-investment through government intervention. Furthermore, SOEs’ managers have the benefitof investment expansion, which yields over-investment through management discretion. Therefore, considering both government intervention and managerial entrenchment of SOEs, it would be more meaningful to understand the current SOE reformation in China by analyzing the inducements of over-investment. Based on the above discus- sion, we pose three questions as follows. (1) What are the impacts of these two forces on SOEs’ investment decisions? (2) How do they interact during the process? Differently put, are they supplementary or complementary to each other? (3) Which one is the dominant force? Is there a difference in managerial entrenchment under heterogeneous levels of government interventions? 3. Research design 3.1. Samples and data The samples used in this study are drawn from A-share firms listed on Shanghai Stock Exchange and Shenzhen Stock Exchange in China during 2003–2010. Our final sam- ples are unbalanced panel data consisting of 7997 firm-year observations after exclud- ing: (1) firms in the financial sector; (2) observations with missing value; (3) cases without sufficient information on the nature of firm ownership and industry; (4) obser- vations in their IPO years, and (5) observations in 2003 because our regression model (1) and (2) uses variables lagged by one period. We winsorize all continuous variables at the 1% and 99% levels. The data on the nature of firm ownership are from the China Center for Economic Research (CCER) database, and the data on managerial power are hand-collected from finance. sina.com.cn. Other firm-specific information is collected from the China Stock Market and Accounting Research (CSMAR) database. China Journal of Accounting Studies 243 3.2. Variable definitions and regression models 3.2.1. Proxy for over-investment Consistent with prior research (e.g., Chen, Hope, Li, & Wang, 2011), we measure over-investment by a positive deviation from expected investment using the model based on Richardson (2006). We estimate optimal investment according to the following regression: Inv ¼ a þ a Growth þ a Lev þ a Cash þ a Age þ a Size t 0 1 t1 2 t1 3 t1 4 t1 5 t1 X X (1) þ a Ret þ a Inv þ Industry þ Year þ e 6 t1 7 t1 where the dependent variable, Inv , is the capital investment in year t. Growth t t–1 represents firm growth opportunities. We use the sale growth rate in year t–1 in the main tests, and Tobin’s Q at the end of year t–1 for robustness tests. Lev , Cash , t–1 t–1 Age , Size , Ret and Inv represent the financial leverage, cash holdings, age of t–1 t–1 t–1 t–1 being listed, firm size, stock return and capital investment at the end of year t–1, respectively. In addition, Industry and Year are dummy variables used to control for the industry and year effects. The residual ε represents deviations from expected investment, and positive deviations are considered as over-investments (OverInv). 3.2.2. Proxy for government intervention Providing the fact that the policy burden that Chinese enterprises bear reflects the direct economic consequences of government intervention in SOEs (Lin & Li, 2008), the pol- icy burden is used as a proxy for government intervention. To implement a compara- tive-advantage defying (CAD) strategy, the Chinese government needs to impose its policy burdens on firms, known as the strategic policy burden and the social policy burden. The former stems from the fact that firms are forced to enter CAD industries or adopt CAD technologies, which were transferred to encourage local economic growth (Qian & Roland, 1998). The latter refers to keeping redundant workers. Since the capital-intensive industries cannot provide enough job opportunities, in order to solve employment problems and maintain social stability, Chinese government expects SOEs to retain excess workers, resulting in a social policy burden (Lin & Li, 2008). Either the strategic policy burden or the social policy burden is a result of the govern- ment strategy of CAD. Under the government intervention, firms invest in projects that should be financed by the government or provide jobs that should be removed. In the trading game between government and firms, stronger government interventions yield heavier policy burdens for firms. Lin, Liu and Zhang (2004) employed the square of the deviation for a firm’s optimal capital–labor ratio determined by economic factor endowments, as a proxy for policy burden. Motivated by their research design, we measure policy burden by using an expanded model based on Zeng and Chen (2006) and Liu, Zhang, Wang, and Wu (2010). We estimate the optimal capital–labor ratio based on the following regression specification. Ci ¼ b þ b Size þ b Lev þ b Roa þ b Growth þ b Tangible t t1 t1 t1 t1 t1 0 1 2 3 4 5 X X X (2) þ Zone þ Industry þ Yearþd where the dependent variable, Ci , is capital-intensity, defined as capital–labor ratio, which is measured as the ratio of a firm’s net value of Property, plant and equipment (PPE) to the number of employees at the end of year t. Size , Lev , Roa and t–1 t–1 t–1 244 Bai and Lian Tangible represent the firm size, financial leverage, return on assets and tangible t–1 assets at the end of year t–1, respectively. Zone is the regional dummy variable. In addition, we control for industry and year effects. The residual δ represents the devia- tion from the optimal capital-intensity determined by economic factor endowments. A positive residual indicates that a firm’s capital-intensity surpasses its optimal capital- intensity, primarily driven by the strategic policy burden. A negative residual indicates that a firm’s capital-intensity is lower than its optimal capital-intensity, primarily driven by social policy burden. We use the absolute value for residual δ as the policy burden (Ovci). 3.2.3. Proxy for managerial entrenchment The agency problem stems from the separation of ownership and control. A manager has discretion over firms’ decisions because he or she ends up with significant control rights of the firm (Shleifer & Vishny, 1997). Consistent with Lu, Wei, and Li (2008) and Quan, Wu, and Wen (2010), we measure managerial entrenchment by using mana- gerial power to reflect managers’ ability to expropriate shareholders or to misallocate firms’ funds. Managers with a stronger managerial power have more discretion over firms’ decisions, and thus are more likely to serve the interests of themselves rather than those of shareholders. Managerial power depends on two aspects: first, how do managers acquire and reinforce their power, which comes from the balance of power between managers and the board members. Second, how are they constrained or moni- tored, which mainly comes from the balance of power between the large and minority shareholders. Therefore, we measure the managerial power from the above two aspects: managerial power structure and ownership structure. The former reflects the position power, which will be strengthened if a manager, in particular, the CEO, is also a member of the board, or even the chairman of the board. The latter reflects the balance of power of ownership. When control rights are concentrated in the hands of a smaller number of shareholders, the managers may have less discretion over the firm, and thus they have less power. 3.3. Baseline regression: government intervention, managerial entrenchment and over-investment Next, in model (3), we examine how government intervention and managerial entrenchment affect over-investment. X X OverInv ¼ c þ c Ovci þ c Power þ c Ovci Power þ Control þ Year þ l t t t t t t 0 1 2 3 (3) where the dependent variable, OverInv, represents over-investment, policy burden (Ovci) is a proxy for government intervention, and managerial power (Power) is a proxy for managerial entrenchment. The coefficients γ and γ denote the correlations of govern- 1 2 ment intervention and managerial entrenchment with over-investment, respectively; the coefficient on the interaction term (Ovci × Power)is γ , which captures the relationship between government intervention and managerial entrenchment and its influence on over-investment. If γ and γ are significantly positive, while γ is significantly negative, 1 2 3 both government intervention and managerial entrenchment potentially lead to over- investment but have a complementary relationship. In comparison, if γ is significantly 3 China Journal of Accounting Studies 245 positive, the government and management potentially have a supplementary relationship, wherein the two conspire to further encroach on external investor interests. The Control comprises a set of control variables that could affect over-investment, including the man- agement expense (Adm), free cash flow (Fcf), cash occupied by related parties (Orecta), executive compensation (Comp), the proportion of independent directors (Indep) and the marketization index (Market). Ang, Cole, and Lin (2000) argue that the management expense ratio is an economic consequence of conflicts between managers and shareholders, which may also result in over-investment; thus, it is unlikely to predict the relation between these two variables. Jensen (1986) argues that empire-building preferences will cause managers to spend essentially all free cash flow on investment projects. The considerations lead to the pre- diction that over-investment will be increasing in free cash flow (Fcf). Xin et al. (2007) argue that ‘other receivables’ used as a vehicle for large shareholder tunneling will make firms more financially constrained, and thus they have to cut the capital invest- ment, which finally make them less likely to over-invest. Therefore, we expect cash occupied by related parties (Orecta) to have a negative impact on a firm’s over-invest- ment. Xin et al. (2007) also find that the existence of regulation on executive compen- sation in Chinese SOEs leads to the execution failure of the compensation contraction between the executives and the governments, making the executives more likely to over-invest. We therefore predict that executive compensation (Comp) would have a negative impact on over-investment. Zhi and Tong (2005) find that independent direc- tors can effectively identify the managers’ earnings management behavior, and Ye, Lu and Zhang (2007) find that independent directors can reduce the funds embezzled by large shareholders. Tang, Luo, and Wang (2005) and Gao, He, and Huang (2006), how- ever, fail to find a significant and negative relation between the proportion of indepen- dent directors on board and the extent of tunneling in a firm. Inconsistencies in empirical evidence and the conclusions of prior research make it hard to predict the relation between the proportion of independent directors on board (Indep) and over- investment (OverInv). Tang et al. (2010) argue that, firms located in places where governments have strong incentives to influence their operations, are more likely to over-invest. Therefore, we expect that the marketization index (Market) would have a negative coefficient. We adjust all continuous variables by subtracting the industry median for the current year, except over-investment (OverInv), policy burden (Ovci) and the marketization index (Market). Detailed variable definitions are provided in Table 1. 4. Descriptive statistics and empirical results 4.1. Variable estimates and descriptive statistics Table 2 reports the regression results based on Model (1). Consistent with prior research (e.g., Xin et al. 2007), the coefficients on Growth , Cash , Size , Ret t–1 t–1 t–1 t–1 and Inv are all positive and significant at a 1% level, and Lev and Age are nega- t–1 t–1 t–1 tively and significantly associated with Inv . According to the regression results of model (1), we obtain 3177 observations in the over-investment sample, including 2108 SOE observations and 1069 non-SOE observations. Table 3 reports the regression results based on model (2). Consistent with prior studies (e.g., Liu et al., 2010), the coefficients on Size and Tangible are positive t–1 t–1 and significant at a 1% level, and Lev is negatively and significantly associated with t–1 246 Bai and Lian Table 1. Definitions of the variables. Variables Definition Inv Capital investment is measured as cash payments for fixed assets, intangible assets, and other long-term assets from the cash flow statement minus receipts from selling these assets, scaled by beginning total assets. OverInv Over-investment is measured as the positive residuals of model (1). Ovci Policy burden is measured as the absolute value of the residuals of model (2). State A dummy variable equals one if the firm is an SOE and zero if the firm is a non-SOE. Structure Managerial power structure is measured as an ordinal dummy variable, which equals one if the CEO is not a member of the board, two if the CEO is also the member of the board, and three if the CEO is also the chairman of the board. Disp Ownership structure is measured as the ratio of the sum of the percentage points of shareholding by the 2nd to the 10th largest shareholders to the percentage points of shareholding by the 1st largest shareholders. Power Managerial power is measured as the sum of Structure and Disp after being standardized. Adm Management expense is measured as the ratio of administrative expenses to total revenue. Growth Firm growth opportunities is measured as Tobin’s Q or Dsales. Tobin’s Q is measured as the sum of market value of tradable shares, book value of non-tradable shares and liabilities divided by book value to total assets. Dsales is measured as the proportion of change in sales. Lev Financial leverage is measured as the ratio of total liabilities to total assets. Cash Cash holdings is measured as the ratio of cash to total assets. Age Age of being listed is measured as the number of years a firm being publicly listed. Size Firm size is measured as the natural logarithm of a firm’s total asset. Ret Stock return is measured as the annual market stock return of a firm. Ci Capital-intensity is measured as the ratio of net value of Property, plant and equipment (PPE) (in million yuan) to the number of employees. Tangible Tangible assets is measured as the ratio of net value of Property, plant and equipment (PPE) to total assets. Roa Return on assets is measured as the ratio of net income to total assets. Fcf Free cash flow is measured as the ratio of the operating cash flow of a firm to total assets. Orecta Cash occupied by related parties is measured as the ratio of other receivables to total assets. Comp Executive compensation is measured as the natural logarithm of cash compensation of the sum of the three highest paid executives. Indep The proportion of independent directors on board. Market A comprehensive index measuring the development of the regional market in which a firm is registered (Fan & Wang, 2010), where higher values indicate greater regional market development. Zone A region-level dummy variable to proxy for the degree of development of China’s regional institutions. Zone equals one if a firm is registered in coastal areas, and zero otherwise. Zone equals one if a firm is registered in central areas, and zero otherwise. Industry The classification of industry follows the CSRC document, Guidance on Listed Firms’ Industries, issued on April, 2001. There are altogether 13 industries coded from A, to M, and 10 subindustries under C. We classify all the listed firms into 22 industries as we treat the 10 subindustries under manufacturing as distinct industries. Year A dummy variable equals one if the firm went public during that year, and zero otherwise. China Journal of Accounting Studies 247 Table 2. Regression results of investment equation. Variables Inv *** Intercept –0.0359 (–3.12) *** Growth 0.0026 t–1 (2.74) *** Lev –0.0171 t–1 (–7.34) *** Cash 0.0328 t–1 (6.11) *** Age –0.0007 t–1 (–4.65) *** Size 0.0032 t–1 (6.04) *** Ret 0.0027 t–1 (3.35) *** Inv 0.5062 t–1 (40.12) Year and Industry Yes Adj-R 0.407 N 7997 * ** *** , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively, for a two-tailed test. T-statistics are reported in parentheses, and are based on Huber-White’s robust standard errors. Ci . The coefficient on Roa is negative, while the coefficient on Growth is positive, t t–1 t–1 but none of them are statistically significant. Table 4 provides the descriptive statistics for main variables. As reported in the table, both the number of over-investment observations (n) and the mean or median over-investment (OverInv) are significantly larger for the SOE group than the corresponding numbers for the non-SOE group. Furthermore, in the SOE group, the policy burden (Ovci) is significantly higher than that in the non-SOE group, while Table 3. Regression results of policy burden equation. Variables Ci *** Intercept –4.6127 (–11.23) *** Size 0.2339 t–1 (12.36) ** Lev –0.1885 t–1 (–2.54) Roa –0.3293 t–1 (–1.39) Growth 0.0268 t–1 (0.64) *** Tangible 1.4645 t–1 (10.91) Year and Industry Yes Adj-R 0.184 N 7997 * ** *** , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively, for a two-tailed test. T-statistics are reported in parentheses, and are based on Huber-White’s robust standard errors. 248 Bai and Lian Table 4. Descriptive statistics for main variables. Test for difference in Full sample SOEs: State=1 Non-SOEs: State=0 mean and median Variables Mean Median Mean Median Mean Median t value z value * * OverInv (n) 0.0388 (3177) 0.0242 (3177) 0.0397 (2108) 0.0246 (2108) 0.0370 (1069) 0.0236 (1069) 0.0027 0.0010 *** *** Ovci 0.6289 0.3061 0.6827 0.3291 0.5232 0.2657 0.1595 0.0634 *** *** Power 0.0000 –0.3559 –0.1952 –0.5553 0.3762 0.0853 –0.5714 –0.6406 *** *** Adm 0.1062 0.0715 0.0933 0.0684 0.1309 0.0773 –0.0376 –0.0089 *** *** Fcf 0.0527 0.0515 0.0553 0.0531 0.0479 0.0475 0.0074 0.0056 *** *** Orecta 0.0406 0.0154 0.0355 0.0142 0.0504 0.0183 –0.0149 –0.0041 *** *** Comp 13.4890 13.5222 13.5110 13.5515 13.4465 13.4568 0.0645 0.0947 *** *** Indep 0.3574 0.3333 0.3550 0.3333 0.3620 0.3333 –0.007 0.0000 * ** *** , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively. The test for mean difference is student’s test, and the test for median difference is Wilcoxon test. Figures in parentheses denote number of the observation for firms with overinvestment. China Journal of Accounting Studies 249 managerial power (Power) is significantly less than that in the non-SOE group. In addi- tion, the management expense (Adm) and cash occupied by related parties (Orecta) are significantly lower for the SOE group than those for the non-SOE group, while free cash flow (Fcf) in the SOE group is significantly higher than that in the non-SOE group. The results from univariate tests indicate that, relative to non-SOEs, SOEs assume more policy burdens due to government intervention, acquire more funds due to government support and have a lower agency cost due to government supervision and restriction of managerial power. From the investment efficiency perspective, govern- ment intervention may cause over-investment due to its social responsibility goals, while also limiting the manager’s over-investment behavior by constraining managerial power. Overall, in China, wherein market competition and corporate governance mech- anisms are not well developed, both the government’s protective and predatory side may coexist in SOEs. On one hand, in order to encourage economic growth and main- tain social stability, the government imposes more policy burdens on SOEs, making their decisions deviate from the goal of maximizing shareholders’ interests and instead, reducing shareholders’ wealth. That’s the grabbing hand of government. On the other hand, by strengthening supervision over managers, the government can limit managerial entrenchment, and therefore protect shareholders’ interests from expropriation by man- agers, and that’s the helping hand of the government. Table 5 reports Pearson and Spearman correlations for the main variables. Both policy burden (Ovci) and managerial power (Power) are significantly and positively correlated with over-investment (OverInv) whether the Spearman or Pearson correlation coefficient is used, indicating that both government intervention and managerial entrenchment can lead to over-investment. The government may be highly motivated to request firms expand investment to achieve social goals, such as promoting local economic growth and maintaining social stability, which results in over-investment. Managers with greater managerial power have greater motivation and capacities for over-investment. Policy burden (Ovci) and managerial power (Power) are significantly and negatively correlated, suggesting a complementary relationship between these two variables. For given interest constraints, government and managers are involved in competitive games in expropriating the outside shareholders. In addition, managerial power (Power) is also significantly and positively correlated with management expenses (Adm) and cash occupied by related parties (Orecta), indicating that the greater the managerial power, the greater tendency to opportunism managers would have. Conversely, the policy burden (Ovci) is significantly and negatively correlated with cash occupied by related parties (Orecta) and management expenses (Adm), implying that government intervention can restrict management opportunistic behavior. Conse- quently, the government, as a regulatory authority, has a restrictive effect on managerial entrenchment. Meanwhile, policy burden (Ovci) is significantly and positively correlated with free cash flow (Fcf), indicating that government intervention can help firms acquire more fund as well. 4.2. Effects of government intervention and managerial entrenchment on over-investment Table 6 reports the regression results for government intervention, managerial entrenchment and over-investment in SOE and non-SOE groups. In SOEs, both the coefficient on policy burdens (Ovci) and that of managerial power (Power) are positive 250 Bai and Lian Table 5. Correlation matrix for main variables. OverInv Ovci Power Adm Fcf Orecta Comp Indep *** * *** *** *** * OverInv 0.0776 0.0322 –0.0462 0.1202 –0.0960 –0.0088 –0.0347 *** *** *** ** *** Ovci 0.0732 –0.0497 –0.0120 0.0402 –0.0256 0.0611 0.0034 ** ** *** *** *** Power 0.0407 –0.0401 0.1103 0.0089 0.0653 0.1047 0.0062 *** ** *** *** *** Adm –0.0237 –0.0578 0.0974 –0.1453 0.4102 –0.1876 0.0002 *** *** *** *** *** *** *** Fcf 0.1347 0.0616 0.0452 –0.0708 –0.1454 0.1079 –0.0512 *** *** *** *** *** *** ** Orecta –0.1639 –0.0902 0.0620 0.2662 –0.1339 –0.2376 –0.0293 *** *** *** *** *** *** Comp 0.0074 0.1098 0.0938 –0.1160 0.1343 –0.1996 0.0612 *** * Indep –0.0104 0.0209 0.0162 0.0079 –0.0515 0.0039 0.0316 * ** *** , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively. This table reports Pearson (above the diagonal) and Spearman (below the diagonal) correlation matrix of the main variables. China Journal of Accounting Studies 251 Table 6. Impacts of government intervention and managerial entrenchment on over-investment. SOEs: State=1 Non-SOEs: State=0 Variables (1) (2) (3) (4) (5) (6) *** *** *** *** *** *** Intercept 0.0482 0.0489 0.0471 0.0490 0.0488 0.0487 (11.57) (11.73) (11.25) (7.91) (7.89) (7.80) *** *** Ovci 0.0031 0.0030 0.0013 0.0012 (3.59) (3.62) (1.08) (1.05) ** *** Power 0.0017 0.0028 0.0012 0.0007 (2.29) (3.57) (1.55) (0.89) *** ** Ovci×Power –0.0014 0.0010 (–2.61) (2.35) Adm 0.0201 0.0176 0.0175 –0.0076 –0.0089 –0.0086 (1.85) (1.60) (1.61) (–1.14) (–1.32) (–1.27) *** *** *** *** *** *** Fcf 0.0463 0.0460 0.0457 0.0575 0.0555 0.0558 (3.43) (3.40) (3.40) (3.51) (3.40) (3.42) *** *** *** Orecta –0.1113 –0.1122 –0.1142 –0.0118 –0.0123 –0.0112 (–7.87) (–8.02) (–8.07) (–0.61) (–0.63) (–0.58) Comp –0.0020 –0.0024 –0.0027 0.0011 0.0007 0.0009 (–1.38) (–1.64) (–1.91) (0.66) (0.44) (0.55) Indep –0.0076 –0.0110 –0.0063 –0.0212 –0.0194 –0.0221 (–0.40) (–0.56) (–0.33) (–0.79) (–0.73) (–0.83) ** ** ** ** ** ** Market –0.0012 –0.0010 –0.0011 –0.0014 –0.0014 –0.0014 (–2.41) (–2.12) (–2.22) (–2.10) (–2.11) (–2.16) Year and Yes Yes Yes Yes Yes Yes Industry Adj-R 0.032 0.025 0.038 0.017 0.018 0.019 N 2108 2108 2108 1069 1069 1069 The table provides the regression results of government intervention and managerial entrenchment on over- investment on SOEs and non-SOEs. The dependent variable OverInv is calculated according to Richardson * ** *** (2006). , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively, for a two-tailed test. T- statistics are reported in parentheses, and are based on Huber-White’s robust standard errors. and statistically significant, while the coefficient on the interaction term (Ovci × Power) is significantly negative. The results indicate that both government intervention and managerial entrenchment can lead the SOEs to over-invest. Meanwhile, these two are complementary to each other in affecting over-investment. However, in non-SOEs, nei- ther the coefficient on policy burden (Ovci) nor that of managerial power (Power)is significant and, in particular, the coefficient on the interaction term (Ovci × Power)is significantly positive. Such a difference suggests that government intervention, manage- rial entrenchment and their interaction effects can be found mainly in SOEs. The gov- ernment can impose policy burdens such as economic growth and social stability on SOEs, resulting in over-investment in SOEs. Meanwhile, as agents, managers may also expropriate outside shareholders through over-investment to pursue their own benefits. However, the counterbalance between government and managers can suppress over- investment. The interaction effects between government intervention and managerial entrenchment on over-investment in SOEs can be explained as follows. On one hand, through supervising the manager and optimizing the protective mechanism for minority shareholders, the government increases the costs of expropriating outside shareholders by managers, thereby restricting such expropriation. On the other hand, managers may resist the predatory behavior of government to pursue their own benefits. 252 Bai and Lian 4.3. Further analysis The above analyses show that in SOEs, both government intervention and managerial entrenchment can lead to over-investment, while these two effects appear to be comple- mentary to each other. One may expect to see which is the dominant force, government intervention or managerial entrenchment? If government intervention dominates mana- gerial entrenchment, then how do heterogeneous levels of government intervention affect managerial entrenchment? Differently put, is there a difference of managerial entrenchment under heterogeneous levels of government interventions? To answer these questions, we examine the interaction effects between government intervention and managerial entrenchment on SOEs’ over-investment under heterogeneous levels of government interventions in this section. 4.3.1. Is government intervention the dominant force leading to over-investment? In this section, we divide the sample firms into groups with high or low government inter- vention by using the median value of policy burden (Ovci), and then examine the interac- tion effects between government intervention and managerial entrenchment on over- investment between SOEs with high government intervention and SOEs with low. We create a dummy variable, Ctrllevel, which is coded as one if the policy burden (Ovci)is greater than the industry median value in SOEs in that year, and zero otherwise. Descriptive statistics are reported in Table 7. It shows that both the mean and median of over-investment (OverInv) are significantly higher for the SOEs with high government intervention than for those with low, while there is no statistical difference for managerial power (Power) between these two groups (SOEs with high versus low government intervention). The results of univariate tests indicate that SOEs under heter- ogeneous levels of government interventions have different over-investment expendi- tures when the managerial power is relatively constant. Hence, government intervention may be the dominant force that affects SOEs’ over-investment. For the SOEs with high government intervention, the government imposes more policy burdens on SOEs, which leads to over-investment. 4.3.2. Different management behavior under heterogeneous levels of government interventions The further univariate results reported in Table 7 suggest that government intervention may be the dominant force affecting SOEs’ over-investment. We subsequently examine Table 7. Descriptive statistics of over-investment and managerial power. High government Low government intervention intervention Test for difference in （Ctrllevel=1） （Ctrllevel=0） mean and median Variables Mean Median Mean Median t value z value *** *** OverInv 0.0431 0.0269 0.0364 0.0229 0.0067 0.0040 (n) (1053) (1053) (1055) (1055) Power -0.1121 -0.4708 -0.1916 -0.5354 0.0795 0.0646 *** *, **, and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively. The test for mean difference is student’s test, and the test for median difference is Wilcoxon test. Figures in parentheses denote number of the observation for firms with overinvestment. China Journal of Accounting Studies 253 the interaction effects between government intervention and managerial entrenchment on SOEs’ over-investment and the behavioral characteristics of managers in response to heterogeneous levels of government interventions. The empirical results are reported in Table 8. The results indicate that for SOEs with high government intervention, the coefficient on policy burden (Ovci)is significantly positive at the 1% level and the coefficient on managerial power (Power) is positive at the 10% level, while the coefficient on the interaction term (Ovci×Power) is significantly negative at the 5% level, suggesting that for SOEs within a strong government intervention environment, government supervision on a manager is greater when the government has stronger incentives to intervene in SOEs’ decisions. A strong control by the government over SOEs constrains managers’ opportunistic behavior, thereby weakening the managers’ discretion over firms. However, managers may resist government intervention as long as they have incentives to maximize their personal benefits of control. For SOEs with low government intervention, policy burden (Ovci) is significantly positive at the 10% level and managerial power (Power) is significantly positive at the 1% level correlation with over-investment, while the interaction term Table 8. Determinants of overinvestment: SOEs with high or low government intervention. SOEs (State=1) High government Low government intervention intervention （Ctrllevel=1） （Ctrllevel=0） Variables （1）（2）（3）（4）（5）（6） *** *** *** *** *** *** Intercept 0.0518 0.0542 0.0512 0.0430 0.0446 0.0415 (8.25) (8.73) (8.11) (7.15) (7.82) (6.84) *** *** * * Ovci 0.0025 0.0023 0.0211 0.0230 (2.64) (2.62) (1.67) (1.84) * *** *** Power 0.0004 0.0022 0.0032 0.0051 (0.34) (1.82) (3.20) (2.78) ** Ovci×Power –0.0013 –0.0115 (–2.34) (–1.22) *** ** ** Adm 0.0105 0.0105 0.0114 0.0301 0.0228 0.0231 (0.53) (0.53) (0.57) (2.69) (2.00) (1.99) ** ** ** *** *** *** Fcf 0.0456 0.0450 0.0458 0.0460 0.0442 0.0450 (2.09) (2.06) (2.10) (2.78) (2.67) (2.74) *** *** *** *** *** *** Orecta –0.1063 –0.1027 –0.1057 –0.1128 –0.1183 –0.1191 (–4.75) (–4.75) (–4.76) (–6.66) (–6.96) (–6.98) * * ** Comp –0.0040 –0.0041 –0.0047 0.0000 –0.0005 –0.0004 (–1.83) (–1.88) (–2.18) (0.03) (–0.28) (–0.24) Indep –0.0306 –0.0371 –0.0292 0.0113 0.0119 0.0139 (–1.06) (–1.28) (–1.01) (0.44) (0.46) (0.54) * * * Market –0.0012 –0.0012 –0.0012 –0.0011 –0.0010 –0.0010 (–1.73) (–1.63) (–1.65) (–1.71) (–1.44) (–1.53) Year and Yes Yes Yes Yes Yes Yes Industry Adj-R 0.029 0.019 0.032 0.025 0.033 0.035 N 1053 1053 1053 1055 1055 1055 The table provides the regression results of interaction effects between government intervention and managerial entrenchment on over-investment with high or low government intervention. The dependent * ** *** variable OverInv is calculated according to Richardson (2006). , , and indicate the 0.1, 0.05, and 0.01 levels of significance, respectively, for a two-tailed test. T-statistics are reported in parentheses, and are based on Huber-White’s robust standard errors. 254 Bai and Lian (Ovci × Power) is negative but not statistically significant, suggesting that the managers have more discretion over investment decisions in SOEs within a weak government intervention environment. With less supervision from the government, managers are more likely to over-invest and tend toward opportunism. Using China as our research setting, our study complements and extends existing research on the economic consequences of ‘the twin agency problems’ proposed by Stulz (2005). In his model, Stulz argues that all the investors risk expropriation by the government and outside investors additionally risk expropriation by those who control firms, whom he calls corporate insiders, since they are sometimes managers and at other times controlling shareholders. In the transitional economy of China, government (officials) has stronger incentives to intervene in SOEs’ decision-making due to the fiscal decentralization and the reform of political promotion mechanisms, while the managers of SOEs end up with the de facto operating decision rights during the process of the Chinese government launching a program that decentralized the managerial deci- sion rights of SOEs from the central government down to the local firm-level. This is the ‘two agency problems’ described by Stulz (2005). However, different from his model, the Chinese government also plays a protective role, by constraining managerial opportunism, in addition to its expropriating role. Consequently, managerial entrench- ment in SOEs is determined by the institutional environment of government reform. It grants large controlling rights to the managers in SOEs, however, when internal governance fails, it can also be a monitor to prevent the managers from making over-investment decisions. 4.4. Robustness tests We conducted a battery of additional tests to check the robustness of our results. First, we used Tobin’s Q as an alternative measure for growth opportunities, instead of sales growth as used before. Xin et al. (2007) argue that Chinese stock markets are less developed relative to other markets, such as the US markets, which makes the Tobin’s Q an inappropriate measure of growth opportunities. To examine whether our results are robust to using Tobin’s Q, we re-do the analysis by using this alternative measure of growth opportunities. Second, following Xin et al. (2007), we categorize firms into three groups based on the residuals derived from model (1) and remove the middle group because these firms, whose unexpected investments are closest to 0 among all firms, are more likely to be affected by measurement error in the investment model (Chen, Hope, Li, & Wang, 2011). Firms in the group with the largest residuals are regarded as firms with over- investment. Third, the panel data sets we use here contain observations on multiple firms in multiple years, so we correct the standard error for clustering of observations by firms as suggested by Petersen (2009). Fourth, we tackle endogeneity. One may ask whether the relationship between government intervention, managerial entrenchment and over-investment is endogenously determined. Following Lin and Li’s(2008) theoretical analysis, we use the ratio of secondary industry production to gross production, the ratio of population employed by secondary industry to total employed persons, as well as GDP per capita, as instrumental variables of policy burden. Fifth, we use negative residuals derived from model (2) as a proxy for policy bur- den (Ovci). In the above analysis, we measure the policy burden by the absolute value China Journal of Accounting Studies 255 of residual from model (2). As Lin and Li (2008) point out, policy burdens consist of both strategic policy burdens and social policy burdens. A strategic policy burden stems from the fact that the firms are forced to enter CAD industries or adopt CAD technologies. Since the capital-intensive industries cannot provide enough job opportunities, in order to solve employment problems and maintain social stability, government needs to retain excess workers, which results in a social policy burden. Therefore, the social policy bur- den is a more direct economic consequence of government intervention, and it may be a more appropriate measure for the incentives of the government to intervene in SOEs. Sixth, the levels of government intervention and the interaction between government intervention and managerial entrenchment are examined both cross-section- ally and over time. In the cross-sectional analyses, we divide the SOE samples into high or low government intervention groups in two different ways. One, following Fan et al. (2012), uses the unemployment rate of the region under the jurisdiction of the local government to measure the regional policy burden, thereby classifying the levels of government intervention environment. Given that social stability is a major policy burden of government, achieving higher unemployment provides government with the stronger incentives to intervene in corporate investments and results in policy burdens at the regional level. We create a dummy variable, Unemrate_dum, which equals one if the unemployment rate of the region under the jurisdiction of the local government is greater than the median value in China in that year, and zero otherwise. The other way is by using the median value of the marketization index by which we divide the SOE samples into high or low government intervention groups. SOEs with high government intervention are firms located in places where the marketization index is less than the median value in that year, while SOEs with low government intervention are firms located in places where the marketization index is greater than the median value in that year. In the time series analysis, we divide the SOE samples into high or low govern- ment intervention groups in a particular period. During the recent financial crisis, the Chinese government was forced to unleash a plan of government stimulus and credit to keep growth growing, including a 4-trillion Yuan fiscal stimulus and a 10-trillion Yuan bank loans stimulus. The stimulus programs were implemented through direct instruc- tions from the government to banks for loan issuance and project financing to maintain economic growth through capital expenditure expansion. In the process, the government directly intervened in bank loan allocations through commands (Bai & Lian, 2012), imposed its objectives on SOEs, and forced SOEs to expand their investment for politi- cal objectives instead of maximizing the value of corporate shareholders. Therefore, we predict that the level of the government intervention in 2008–2010 should be greater than that in 2004–2007. In summary, our results are robust to alternative measures of growth opportunities, over-investment, policy burden, government intervention and the levels of government interventions, in order to tackle endogeneity, and to correct the standard error for clus- tering of observations by firms. For the sake of brevity, we do not tabulate the results for the additional tests. 5. Conclusions During economic restructuring, fiscal decentralization and SOE reformation result in excessive control of local governments over SOEs and provide incentives to corporate managers to pursue greater authority and controls. Government intervention and managerial entrenchment have become two major drivers for SOEs’ over-investments. 256 Bai and Lian Government may force firms to carry out policy burdens, such as economic develop- ment and employment, leading to SOEs’ over-investment. Alternatively, empowered managers can also expropriate substantial gains from SOEs, resulting in over-invest- ment for their own personal benefits. Moreover, the Chinese government also plays a protective role (‘helping hand’) by constraining managerial opportunism in SOEs, in addition to its role of the ‘grabbing hand’, and in turn managers will resist the monitoring from the government. Our empirical results reveal that, in SOEs, both the policy burden from government intervention and managerial rent-seeking due to managerial entrenchment can lead to over-investment, and these two forces are complementary to each other. More powerful government intervention can suppress managerial entrenchment, and in turn managers have more discretion over firms’ decisions, whose behavior tends toward opportunism where government intervention is weak. Our findings raise an interesting question for further study. Should the Chinese government only play the role as a predator to expropriate resources from these firms, given the government’s incentives to intervene in SOEs for policy burdens and political promotion? If this conjecture is confirmed, one would have to attribute the remarkable economic growth in China to non-SOEs. Our results show that in a transitional econ- omy such as China, where corporate governance and market economy mechanisms are not well developed, government can be conducive to reducing management agency cost through monitoring the managers, ensuring that contracts are fulfilled and partially pro- tecting outside investor interests. Thus, during the market-oriented reform process in China, where the investor protection systems are significantly less developed than most of the countries in the world (Allen, Qian, & Qian, 2005), a gradual reform strategy of deregulation should be persistently implemented as it can retain the alternative systems of informal investor protection that the government brings. Acknowledgements The authors acknowledge financial support from the National Natural Science Foundation of China (No. 71262007), the National Social Science Foundation of China (No. 11XGL002), the Humanistic and Social Science Foundation of the Ministry of Education of China (No. 10YJC630002; No. 09YJC630160), the Social Science Foundation of Xinjiang Production and Construction Corps (No. 10BTYB12) and the Key Research Center for Social Sciences of Committee of Education of Xinjiang (No. XJEDU020112C03). They also thank the Talents Boosting Program for the New Century of the Ministry of Education of China and the National Accounting Talents (Reserve) Program of the Ministry of Finance of China. The authors would like to offer their most sincere thanks to three anonymous reviewers and the editors, Jason Xiao and Tong Yu, for their insightful comments. In addition, the authors specially thank Dr Liang Han, at University of Surrey, UK, for his valuable advice and proofreading. The authors would take full responsibility for the paper. Notes 1. The international data of gross capital formation (% of GDP) is collected from the World Bank (http://data.worldbank.org/indicator/NE.GDI.TOTL.ZS). Other East Asian countries and areas are Cambodia, Hong Kong (PRC), Indonesia, Japan, Lao PDR, Malaysia, Mongolia, South Korea, Singapore, Thailand, and Vietnam, where the data of gross capital formation (% of GDP) are available. 2. Consistent with their findings, Zhang (2003) finds that the investment was efficiently and largely reaped through the rural industrialization and proliferation of small firms in non-state sectors. 3. According to the hypothesis of ‘grabbing hands’ of government, Stulz (2005) came up with the twin agency problems. The core part of his point is that the owners of a firm bear the China Journal of Accounting Studies 257 risks of state expropriation, and outside investors take the risk of expropriation by the insiders, who often refer to controlling shareholders and managers. Specifically, in order to maximize their own interests, insiders would acquire private benefits by expropriating out- side investors, which can produce ‘the agency problem of insider discretion’. At the same time, government rulers expropriate investors by virtue of the authority of government for their own sake, and then ‘the agency problem of state ruler discretion’ arises. 4. Government may also intervene in non-SOEs, but compared with the SOEs, both the possibility and the intensity of the intervention may be lower (Tian, 2005). 5. Because the estimation models of over-investment (OverInv) and policy burden (Ovci) have controlled for the industry and year effects, it can cause repetitive control if we still control for these effects. The marketization index is not affected by industry. Therefore, all continu- ous variables except for OverInv, Ovci and Market are adjusted by using their industry median value. We thank Professor Xin Zhang, at the School of Management, Fudan Univer- sity, for his helpful comments. 6. The standardized equation for variable x is: (x–m/sd, and m stands for the mean of variable x, while sd is the standard deviation. 7. The most updated Marketization Index covers all of the provinces from 1997–2007, so we match the cases after 2008 with the index of 2007. 8. The data are collected from The Statistical Data of New China for Fifty Years. We use statistical data in 1985, because data under some jurisdictions are not available until 1985. In addition, we exclude firms registered in Chongqing, as it was established as a municipality in 9. The unemployment rate data is obtained from www.drcnet.com.cn. 10. Shleifer and Vishny (1994) argue that when the government (officials) controls firms, managers may even use bribes to convince them not to push firms to pursue political objectives, which can be regarded as a means deployed by the managers to resist the intervention by the government (officials). References Allen, F., Qian, J., & Qian, M. (2005). Law, finance, and economic growth in China. Journal of Financial Economics, 77,57–116. Ang, J. S., Cole, R. A., & Lin, J. W. (2000). Agency costs and ownership structure. Journal of Finance, 1,81–106. Bai, J., & Lian, L. S. (2012). Difference of capital allocation: Ownership discrimination or endowments deficiencies? Management World (Chinese), 6,30–42. Blanchard, O., & Shleifer, A. (2001). Federalism with and without Political Centralization: China Versus Russia. IMF Staff Papers, 171–179. Bushman, R., & Smith, A. (2001). Financial accounting information and corporate governance. Journal of Accounting and Economics, 31, 237–333. Chen, D. Q., & Li, S. F. (2012). Government governance, ownership preference and capital investment. Nankai Business Review (Chinese), 1,43–53. Chen, S., Sun, Z., Tang, S., & Wu, D. (2011). Government intervention and investment efficiency: Evidence from China. Journal of Corporate Finance, 17, 259–271. Chen, F., Hope, O. K., Li, Q., & Wang, X. (2011). Financial reporting quality and investment efficiency of private firms in emerging markets. The Accounting Review, 86, 1255–1288. Cheng, Z. L., Xia, X. P., & Yu, M. G. (2008). The government intervention, the pyramidal structure, the investment of local state-owned companies. Management World (Chinese), 9,37–47. Dollar, D. & Wei, S. J. (2007). Das (wasted) Kapital: Firm ownership and investment efficiency in China. NBER working paper. Fan, G., & Wang, X. (2010). NERI index of Marketization of China’s Provinces. Beijing: Economics Science Press (Chinese). Fan, J. P., Wong, T. J., & Zhang, T. (2007). Politically connected CEOs, corporate governance, and Post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics, 84, 330–357. Fan, J. P. Wong, T. J., & Zhang, T. (2012). Institutions and organizational structure: The case of state-owned corporate pyramids. Journal of Law, Economics, and Organization. doi:10.1093/ jleo/ews028. 258 Bai and Lian Fazzari, S. M., Hubbard, R. G., & Petersen, B. C. (1988). Financing constraints and corporate investment. Brookings Papers on Economic Activity, 1988, 141–206. Frye, T., & Shleifer, A. (1997). The invisible hand and the grabbing hand. American Economic Review, 87, 354–358. Gao, L., He, S. H., & Huang, Z. Z. (2006). Corporate governance and tunneling. China Economic Quarterly (Chinese), 3, 1157–1178. Jensen, M. (1986). Agency costs of free cash flow, corporate finance, and takeovers. American Economic Review, 76, 323–329. Jensen, M. (1993). The modern industrial revolution, exit, and the failure of internal control systems. Journal of Finance, 48, 831–880. Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics, 3, 305–360. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (1998). Law and finance. Journal of Political Economy, 106, 1113–1155. Li, H., & Zhou, A. (2005). Political turnover and economic performance: The incentive role of personnel control in China. Journal of Public Economic, 89, 1743–1762. Lin, J. Y., & Li, Z. (2008). Policy burden, privatization and soft budget constraint. Journal of Comparative Economics, 36,90–102. Lin, Y. F., Liu, M. X., & Zhang, Q. (2004). Policy burden and enterprise’s soft budgetary binding: A case study from China. Management World (Chinese), 8,81–89. Liu, H. L., Zhang, M., Wang, Y. P., & Wu, L. S. (2010). Political connections, compensation incen- tive, and employee allocation efficiency. Economic Research Journal (Chinese), 9, 109–121. Lu, R., Wei, M. H., & Li, W. J. (2008). Managerial power, perquisite consumption and performance of property right: Evidence from Chinese listed companies. Nankai Business Review (Chinese), 5,85–92. Macroeconomic Study Group, CCER, Peking University. (2004). Property rights, inefficient investment and deflation. Economic Research Journal (Chinese), 9,26–35. Modigliani, F., & Miller, M. H. (1958). The cost of capital, corporation finance and the theory of investment. American Economic Review, 48, 261–297. National Bureau of Statistics of China. (2011). Bulletin of the verification of gross domestic product (GDP) in Year 2010. http://www.stats.gov.cn/ Oi, J. C. (1992). Fiscal reform and the economic foundations of local state corporatism in China. World Politics, 45,99–126. Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies, 22, 435–480. Qian, Y., & Weingast, B. R. (1997). Federalism as a commitment to preserving market incentives. Journal of Economic Perspectives, 11,83–92. Qian, Y., & Roland, G. (1998). Federalism and the soft budget constraint. American Economic Review, 1143–1162. Quan, X. F., Wu, S. N., & Wen, F. (2010). Managerial power, private income and compensation rigging. Economic Research Journal (Chinese), 11,73–87. Richardson, S. (2006). Over-investment of free cash flow. Review of Accounting Studies, 11, 159–189. Roll, R. (1986). The hubris hypothesis of corporate takeovers. The Journal of Business, 59, 197–216. Shleifer, A., & Vishny, R. W. (1986). Large shareholders and corporate control. Journal of Political Economy, 94, 461–488. Shleifer, A., & Vishny, R. (1989). Management entrenchment: The case of manager-specific investments. Journal of Financial Economics, 25,23–139. Shleifer, A., & Vishny, R. W. (1994). Politicians and firms. The Quarterly Journal of Economics, 109, 995–1025. Shleifer, A., & Vishny, R. W. (1997). A survey of corporate governance. Journal of Finance, 52, 737–783. Stein, J. C. (2003). Agency, information and corporate investment, in G. Constantinides, M. Harris, and R. Stulz (eds), Handbook of the economics of finance. New York: Elsevier/ North-Holland, pp. 111–65. Stulz, R. (1990). Managerial discretion and optimal financing policies. Journal of Financial Economics, 26,3–28. Stulz, R. (2005). The limits of financial globalization. Journal of Finance, 60, 595–1638. China Journal of Accounting Studies 259 Sun, Q., & Tong, W. H. (2003). China share issue privatization: The extent of its success. Journal of Financial Economics, 70, 183–222. Tang, Q. Q., Luo, D. L., & Wang, L. (2005). Controlling shareholders’ tunneling and resistant powers: Evidence from Chinese stock market. China Accounting Review (Chinese), 1,63–86. Tang, X. S., Zhou, X. S., & Ma, R. J. (2007). Empirical research on over-investment behavior and its restriction systems in China’s listed companies. Accounting Research Journal (Chinese), 8,44–52. Tang, X. S., Zhou, X. S., & Ma, R. J. (2010). Government interventions, GDP growth, and local SOE overinvestment. Journal of Financial Research (Chinese), 8,33–48. Tian, L. H. (2005). The national ownership, the budgetary soft control, and the leverage control over China’s listed companies. Management World (Chinese), 7, 123–128. Wei, M. H., & J. H. Liu. (2007). SOEs’ dividend distribution, governing factors and over investment. Management World (Chinese),4, 88–95. Wu, L. S. (2009). State ownership, preferential tax, and corporate tax burdens. Economic Research Journal (Chinese), 10, 109–120. Xia, L., & Chen, X. Y. (2007). Marketization, SOE reform strategy, and endogenously deter- mined corporate governance structure. Economic Research Journal (Chinese), 7,82–95. Xin, Q. Q., Lin, B., & Wang, Y. C. (2007). Government control, executive compensation and capital investment. Economic Research Journal (Chinese), 8,110–122. Ye, K. T., Lu, Z. F., & Zhang, Z. H. (2007). Can independent directors deter the tunneling of large shareholders? Economic Research Journal (Chinese), 4, 101–111. Zeng, Q. S., & Chen, X. Y. (2006). State stockholder, excessive employment and labor cost. Economic Research Journal (Chinese), 5,74–86. Zhang, J. (2003). Investment, investment efficiency, and economic growth in China. Journal of Asian Economics, 14, 713–734. Zhang, H. H., & Wang, Z. J. (2010). Government intervention, government object and stated- owned listed companies’ overinvestment. Nankai Business Review (Chinese), 13, 101–108. Zhang, M., Wu, L. S., & Wang, Y. P. (2010). State ownership, firm performance, and firm investment. Journal of Financial Research (Chinese), 12,115–130. Zhi, X. Q., & Tong, P. (2005). Earnings management, corporate control transfer and independent directors’ turnover. Management World (Chinese), 11, 137–144. Zhong, H., & Y., Ran, M. S., & Wen, S. S. (2010). Government intervention, insider control and corporate investment. Management World (Chinese), 7,98–108.
Journal
China Journal of Accounting Studies
– Taylor & Francis
Published: Dec 1, 2013
Keywords: government intervention; managerial entrenchment; over-investment
