Privatization with ‘Vested Interests’ in China

Privatization with ‘Vested Interests’ in China Abstract Vested interests have been blamed for resisting the reform of state-owned enterprises (SOEs) in China. Yet, various interest groups have heterogeneous interests in privatization. Using both firm- and provincial-level data, we find that SOE managers and local bureaucrats—two key players of privatization—have contingent, rather than vested, interests in privatization, depending partly on their political connections with the central government. On the one hand, political connections motivate SOE managers to privatize more state ownership while retaining managerial control. On the other hand, central connections discourage provincial leaders from using privatization to boost their economic performance. These results shed light on the conditions under which China implements its economic reforms. With the increasing embeddedness and declining autonomy for policymakers, the once well-performing developmental state models now face serious challenges as politically powerful interest groups can manipulate economic reforms for their own purposes rather than for structural transformation. Success in implementing economic reforms depends on a government’s capacity and skill in generating political support and holding off the opposition (Haggard and Webb, 1993). China’s impressive performance in the early stages of reforms can be attributed to two key institutional arrangements. One is its political institutions that allow the government to have ‘embedded autonomy’ (Evans, 1995), which made it easier for policymakers to make difficult decisions without being hampered by interest groups or the need to form cumbersome political coalitions. The other is economic decentralization, in which local governments are highly motivated to engage in competition for growth-enhancing reforms in pursuit of career advancement (Montinola et al., 1995; Xu, 2011). These two factors are inadequate, however, when we seek to account for the complex dynamics of China’s recent experience of economic reforms. The policy-making process, which was characterized by ‘fragmented authoritarianism’ (Lieberthal and Oksenberg, 1988), has become more fragmented and less authoritarian with the increased influence of vested interests (Pei, 2006; Shih, 2007; Landry, 2008; Mertha, 2009). As the World Bank’s report China 2030 puts it, ‘the close links between the government, big banks, and state enterprises have created vested interests that inhibit reforms and contribute to continued ad hoc state interventions in the economy (World Bank 2012)’. Even China’s official Xinhua News Agency (2013) pointed out that ‘some beneficiaries of reform have started to oppose further changes in the country, becoming “powerful vested interests” that obstruct China’s new reforms’. Privatization is an area of reform that has been deeply decentralized and influenced by local governments with the active involvement of managers of state-owned enterprises (SOEs). Vested interests have been blamed for resisting SOE reform. Yet, the reform has been characterized by rapid privatization of ownership but inconsistent policy goals. On the one hand, since 2000 China has been the global leader in privatization transactions with numerous large-scale mergers, management buyouts and takeovers of small SOEs by private firms.1 On the other hand, concerns about rising state capitalism abound as the Chinese government has exerted more effective control on SOEs and increasingly relied on them to pursue various policy goals (Huang, 2008; Bremmer, 2010). We argue that SOE managers and local government leaders—two key players of privatization—are self-seeking agencies with heterogeneous interests in privatization. SOEs are not entrenched interest groups that consistently oppose privatization, and local government leaders are not reformers that always favor privatization. Whether these two politically powerful groups support or oppose privatization depends partly on their connections with the central government. On the one hand, political connections reassure SOE managers that their benefits and privileges will be preserved, therefore enhancing their incentive to push for privatization in pursuit of performance-based personal rewards. On the other hand, political ties with the central government reduce the pressure from provincial leaders to achieve strong economic growth. It gives them more leeway to put non-economic goals ahead of economic performance goals, therefore reducing their incentive to pursue privatization schemes. We test the effects of political connections on the progress of privatization using both firm- and provincial-level data analyses. At the firm level, we find that centrally connected firms are less likely to be controlled by state-equity holders and have more dispersed ownership structures, indicating that connected SOEs tend to pursue more aggressive privatization schemes. The results are robust after we use the entropy balancing method to mitigate potential selection biases. At the provincial level, we find that where leaders have close ties with the central government, SOEs make up a greater share of listed companies. These results suggest that SOE managers and provincial leaders have contingent, rather than vested, interests in privatization, shaped by their political connections with the central government. This study makes three major contributions. First, it contributes to the literature on the political economy of economic reform. Previous studies suggest that the perceived distributional effects of economic reform will form organized interests for or against economic reforms (Fernandez and Rodrik, 1991; Przeworski, 1991; Haggard and Kaufmann, 1992; Rodrik, 1996; Hellman, 1998; Schamis, 1999). But given the uncertain consequences of economic reform and the government’s commitment problem, collective interests turn out to be more fluid and nuanced than standard interest-based models would predict. This study suggests that political connections, as a partial solution to the government’s commitment problem, are an important factor that shapes the preferences of interest groups concerning economic reforms. Second, it provides a new perspective for understanding business–government relations in authoritarian regimes. Unlike the early developmental state literature, which emphasizes an authoritarian government’s role as a helping hand, and the bureaucratic authoritarianism literature, which focuses on the predatory role of government, this study suggests that a web of business–government ties enhances state influence while constraining state autonomy in economic policy making and implementation. In particular, it implies that the pressure for diminishing the state’s interventive role does not just come from private sector or social groups, but could come from SOEs whose agendas conflict with the state’s original plan. Third, it provides empirical evidence for understanding the consequences of China’s state capitalism. Rather than developing a mutually reinforcing relationship between the state and SOEs, state capitalism seems to create interest groups whose agendas often conflict with the state’s objectives. On the one hand, the state continues to maintain widespread influence despite the gradual process of privatization. On the other hand, economic, rather than political and social, motives have increasingly become the key factor shaping the activities of partially privatized SOEs. The seemingly almighty state capitalism has not developed a successful integration of power and wealth. It is not just because SOEs are less efficient than private firms (Lardy, 2014), but also because the divergence of interests between the state and SOEs may be greater than anticipated. This article proceeds as follows. The next section reviews China’s experience of partial privatization. The third section discusses how political connections may shape the preferences of SOE managers and local politicians concerning privatization and presents two testable hypotheses. The fourth section introduces the datasets and variables. In the fifth section, we present the statistical results from both firm- and provincial-level analyses and run various robustness checks. We conclude in the last section. 1. Road to partial privatization in China Reforms of SOEs in China have a complex history that has been characterized by vague and ambiguous initiatives from the central government and decentralized privatization schemes across regions. SOEs have been making the transition to mixed ownership structures since the late 1980s. In the early 1990s, the central government introduced the split-share structure, allowing SOEs to convert a small portion of state ownership into public ownership while remaining in control of the majority state ownership.2 Through this policy, SOEs can issue roughly equal amounts of state, legal person and public shares. Meanwhile, under the slogan ‘zhuada fangxiao’ (grasp the big, let go of the small), Beijing gave local governments the green light to privatize most of the industries that had little national security or fiscal importance. Eighty percent of the small and medium-sized SOEs were privatized (Johnson and Beiman, 2007). For those still controlled by the state, the restriction on the conversion of non-tradable state and legal-person shares to public ones was eventually lifted in 2005, with a grace period of up to three years.3 Because the privatization campaign was carried out without a clear national legal framework, local governments have a great deal of autonomy in deciding their own paces, and they use a variety of methods to respond to local needs. The primary method is direct sales of state ownership to insiders through management buyouts or to outside private owners; this accounts for about 70% of privatization cases in China. Other approaches include public offerings, joint ventures, leasing and employee shareholding (Gan et al., 2012). Local governments also allowed SOEs to negotiate privatization plans with public shareholders. State shareholders can offer extra shares and generous dividend payouts in exchange for public investors’ approval of the equal trading rights of non-tradable shares (Liao et al., 2014). By the end of 2011, 72% of central SOEs and their subsidiaries had been corporatized and almost all local SOEs had been restructured into shareholding companies with multiple shareholders (SASAC, 2012, p. 31). A new round of SOE reforms was unveiled in November 2013, granting a greater space for local governments and SOEs to find their own best practices to establish mixed ownership structures and adopt a more radical ‘wealth management’ approach to privatization. Large-scale partial privatization led to the rapid decline of SOEs’ role in the country’s economy. Between 2000 and 2012, SOEs’ share of industrial assets declined from 47 to 20%, industrial revenues 29 to 12%, industrial profits 19 to 9% and urban employment 35 to 18% (National Statistical Bureau, 2016). The ownership structure of SOEs has also changed profoundly. Between 1992 and 2012, the total assets of SOEs increased 20 times, but their share of state assets dropped from 49 to 28% (Ministry of Finance, 2008, 2013). Simply focusing on the ownership structure, however, reveals little about the actual degree of autonomy SOE managers have over managerial decision-making and influence the state exerts over the firm. There are significant discrepancies between state ownership and actual control of firms. On the one hand, with the massive delegation of managerial discretion and sales of state ownership shares to entrenched insiders, SOE managers enjoy greater managerial autonomy than one might expect given state ownership (Milhaupt and Zheng, 2016). SOE managers are not just granted a great deal of managerial autonomy, they are also compensated with high-powered incentives. Particularly for managers of SOEs controlled by the central government (hereafter central SOEs), their salaries are explicitly tied to the performance evaluation of their companies.4 They could receive bonuses as high as three times their base salaries if their companies receive an ‘A’ in performance rating (SASAC, 2003). The average compensation for a central SOE manager, for example, was 40% higher than the average level for other top executives in 2013 (Beijing Youth Daily, 2014). On the other hand, under the party–state structure, the state can exert greater influence on partially privatized SOEs than what the ownership structure would suggest, reflecting the distinct characteristics of China’s state capitalism (Lin and Milhaupt, 2013). With the creation of the State-owned Asset Supervision and Administration Commission (SASAC) in 2003, the government has reinforced the elite status of mega-SOEs. In the shadow of party control, SASAC executes the power of selection and compensation of top SOE managers, which ensures the state retains substantial control in key corporate decisions. For partially privatized SOEs, minority private shareholders have no effective right to select executive management teams. Their sole right is to receive dividends at the discretion of state shareholders (Yosuf et al., 2006, p. 90). ‘The SOE reform is now a half-way business’, an investment strategist warned, ‘local governments or central SOEs appear willing to undertake reforms which are beneficial to the vested interests, but seem reluctant to do anything more than that’ (Chan, 2014). 2. Vested interests in privatization Earlier studies on the politics of economic reform tend to portray vested interests as organized groups opposing economic liberalization. The success of economic reform hinges on the ability of the government to mobilize political support and overcome opposition from vested interests (Przeworski, 1991; Haggard and Webb, 1993; Hellman, 1998; Denisova et al., 2009). In the early stages of transition, entrenched winners of initial economic reform, primarily SOEs or former SOEs, are motivated—and will have the power to—block future reform to generate concentrated rents for themselves while imposing high social costs (Hellman, 1998; Shleifer and Treisman, 2000). The power of vested interests lies in their ‘formal ties to the state, legacies of previous state ownership, and more frequent interactions with public officials’. In the later stages of economic transition, however, new private firms may organize to use state capture as a substitute to compete against incumbent firms for political influence (Hellman et al., 2003). This analysis of organized interest groups has been widely applied to studies of economic reforms and political transitions in authoritarian regimes (Pei, 2006; Shih, 2007; Malesky and Taussig, 2008; Pepinsky, 2009). In all these studies, vested interests are considered powerful, and sometimes coherently organized, groups that prefer the status quo to reforms. But studies on business–government relations in Latin America suggest that business interests are more fluid, ambiguous and even conflicted, and are affected by reform packages and firm-level characteristics (Murillo, 2002; Schneider, 2004). Even groups that end up benefiting from the reform may be initially in favor of the status quo if they fear that the reform will make them worse off (Fernandez and Rodrik, 1991; Przeworski, 1991). Given this nuanced portrait of business interests, it is prudent to say that vested interests are not necessarily the enemies of reforms, but under what conditions can vested interests be coopted to support policy changes? The cooptation of vested interests often occurs through the creation of new rents to replace old ones or compensating transfers to the losers from reform (Roland, 2002; Shleifer and Treisman, 2000). Vested interests may become reformers if they expect that disruptive developments would make them worse off, or if they believe that institutional changes will be beneficial and will not involve downside risk (Moe, 2015). In regards to privatization, SOE managers represent a key vested interest group who would oppose privatization as they are concerned about losing their old rents—privileged political statuses and financial subsidies, but their position as anti-privatization stakeholders may change if they expect to receive new rents which may give them additional wealth or political power. The effectiveness of the cooptation strategy, however, is contingent on the government’s ability to make a credible commitment (Roland, 2002). To reduce opposition from SOE managers, the government needs to secure their position as the residual claimant after restructuring SOEs. As long as SOE managers are assured that their policy privileges will continue after partial privatization, they would have little incentive to block the reform. Rather, high-powered incentives can encourage SOE managers to shed more unproductive assets through partial privatization. Thus, partial privatization may become a favorable choice because it could provide SOE managers with a significant ownership stake in the firm and reduced outside government intervention without losing their privileged status (Walder, 2011). Without such commitment from the government, SOE managers will behave like vested interests as expected: they are likely to block the privatization scheme in the first place. However, there is another commitment problem in the process of partial privatization. Perotti (1995) argues that if private investors are concerned about post-privatization government interference, they will be less interested in buying government shares or they will request upfront discounts as compensation for risk. Therefore, if the government could not credibly commit to withdrawing its intervention from corporate governance, SOE reforms are likely to fail (Qian, 2003). These two commitment problems are mutually exclusive, however. If the government makes a non-interference commitment to private investors, it will not be able to guarantee benefits to SOE managers because the government has little control over profit distribution in privatized firms. If the government makes commitment to SOE managers, it will have to intervene in the privatized firms to secure compensating transfers. Thus, the government faces a dilemma in the process of privatization. A credible commitment to SOE managers will violate a commitment to private investors, and vice versa. Because no institutional arrangement could possibly resolve the commitment dilemma faced by the government, crony arrangements may become a partial solution, particularly in authoritarian regimes where the government has inherently weak commitment power in the absence of institutional constraints (Haber et al., 2003). Having patrons or representatives in the government not only helps an interest group influence general policy making, but also allows the government to guarantee that the property rights of a subset of asset holders will be protected. Recent studies show that political connections enable powerful firms to gain greater access to particularistic benefits and boost their market values (Hellman and Schankerman, 2000; Faccio, 2006; Faccio et al., 2006; Li et al., 2008; Malesky and Taussig, 2008). Connected SOE managers also receive more personal benefits than those in unconnected firms (Hung et al., 2012). A politically connected SOE manager will be more likely to accept policy changes, if political connections enhance the credibility of the government’s commitment to post-privatization compensating transfer. They may even support privatization because it is in their interest to obtain more management autonomy and ‘depoliticize’ the firm. This discussion yields our first hypothesis. H1: Politically connected SOE managers pursue more aggressive privatization schemes. While SOE managers appear to be vested interests with clear-cut preferences in opposing privatization, local government leaders’ interest in privatization is more ambiguous. On the one hand, controlling SOEs gives local government leaders freedom to use SOEs as critical policy tools to channel public investments and maintain employment, but these benefits depend on the interaction between governments, financial institutions and enterprise managers (Brandt et al., 2005). An increase in market competitiveness, on the other hand, reduces the advantages of state ownership. When the SOE’s budget becomes harder, local leaders may find controlling state ownership less valuable compared to the benefits received from selling the firm. They may be motivated to think SOEs should be privatized if they expect that privatization can increase tax revenue and enhance enterprise efficiency (Boycko et al., 1996; Liu et al., 2006). Although private firms tend to be more efficient and productive than SOEs, the effect of privatization on firm productivity and economic growth is inconclusive in developing countries and transition economies (Clarke and Pitelis, 2005; Megginson, 2005; Bortolotti and Milella, 2006; Estrin et al., 2009). Given the uncertain economic benefits of privatization, whether and how to privatize are essentially the result of political calculations. It is most likely to occur when a local politician cannot obtain large benefits from public firms any more (Shleifer and Vishny, 1994). In democracies, politicians’ motives for privatization depend on their ability to extract political benefits and voting support from privatization (Biais and Perotti, 2002; Dinc and Gupta, 2011). In authoritarian regimes, local politicians also need to survive political competition, but they are, first and foremost, subject to upward rather than downward accountability. In other words, local politicians’ primary motive is to please their superiors rather than their constituencies. In China, where the political regime is characterized by ‘regionally decentralized authoritarianism’ that combines centralized personnel control and a regionally decentralized economic system (Xu, 2011), career-oriented local leaders are motivated to demonstrate political loyalty to gain political credits and deliver strong economic performance to gain economic credits. Jia et al. (2015) find that political connections increase the likelihood of promotion for local leaders, but political connections alone cannot guarantee local leaders’ promotion unless they can deliver strong economic performance. While connections and performance are complements in the promotion of provincial leaders, unconnected provincial leaders will have to rely more on delivering strong economic performance to increase their odds of promotion. The relative importance of political credits and economic credits vary at different levels of government. For leaders at lower-level (county or township) governments, economic performance is a strong predictor of their promotion (Whiting, 2004; Guo 2007; Landry 2008). For provincial government leaders, political factions or connections play a more important role (Shih et al., 2012). Similarly, Bai et al. (2000) argue that higher-level bureaucrats tend to be more concerned about the cost of privatization, such as layoffs of surplus workers and the underprovision of social welfare, whereas lower-level bureaucrats are more likely to pursue the benefits of privatization, such as enhanced profitability and reduced fiscal burdens. Thus, local leaders take both political and economic credits into consideration when implementing privatization plans. They need to find a policy more beneficial for their career prospects than simply pursuing SOE profit maximization (Holz, 2011). Their preferences on privatization may depend on how close they are with the central government.5 We expect that provincial leaders with close ties to the central government will be less likely to pursue privatization because they tend to put non-economic goals (e.g., job creation and social welfare provision) ahead of economic goals whereas those without close ties are more likely to pursue privatization to promote local economic growth. Thus, we have the second hypothesis: H2: Provincial leaders with close ties with the central government pursue less aggressive privatization schemes. 3. Data and empirical strategy To investigate the impact of vested interests and local officials on privatization, we use two datasets. We first use a firm-level dataset to estimate the effect of firm executives’ political connections on firms’ privatization outcomes. Then we use an aggregated dataset to estimate the effect of local leaders’ political connections on the progress of privatization at the provincial level. 3.1 Firm-level analysis The firm-level dataset includes firms that were publicly listed on the main boards of the Shanghai and Shenzhen Stock Exchanges between 2003 and 2009 (ranging from 1285 to 2457). It was assembled from three sources: the China Stock Market and Accounting Research (CSMAR) database for accounting and financial information, the Sinofin database compiled by the China Center for Economic Research (CCER) for information on ownership structure and major shareholders and the annual reports of listed firms for biographical information of firms’ top executives (chairpersons and general managers). In instances where biographical information was not detailed enough in the annual reports, we performed an Internet search to collect company executives’ career information (e.g. bureaucratic or legislative experiences). In addition, all financial variables have been winsorized at the 0.1% and 99.9% levels to minimize the impact of outliers. The panel is unbalanced due to the timing of listing, mergers and acquisitions and missing information for some financial variables. 3.2 Dependent variable Our primary interest is SOEs, but the definition of SOE has evolved over time. Traditional SOEs only refer to enterprises directly owned by the state or state agencies, but the Chinese government has broadened the scope of SOEs to state-invested enterprises (guojia chuzi qiye), which includes those indirectly controlled by the state.6 Since our database does not have information on indirect shareholding, we focus on the changes in types and structures of direct controlling shareholders. There are three major types of shareholder: state, legal person and public shareholders. State shareholders hold state shares (guojia gu) or state-owned legal-person shares (guoyou faren gu).7 Legal-person shareholders hold all other types of non-tradable shares (faren gu). Public shareholders hold all types of publicly tradable shares (gongzhong gu). State ownership. Our first dependent variable is a binary variable that equals 1 if the controlling shareholder holds state equities and 0 otherwise. In 2003, 68% of controlling shareholders of listed firms held state equities. In 2012, the trend changed with only 14% of controlling shareholders holding state equities. State share. To be sure, classifying firms based on whether the majority or dominant owner is state cannot neatly separate firms into state and private domains as many firms have always had hybrid forms of ownership in which state units are sometimes the underlying owner (Lardy, 2014). Therefore, a change in shareholding forms from state equity to non-state equity does not necessarily mean that the state has relinquished control. It could simply be the result of regulatory change that allows state owners to convert their non-tradable state equities into tradable public equities. Thus, we use another variable, state share, which is defined as the proportion of equity owned by the controlling state shareholder, to capture changes in the degree of ownership concentration. During the period 2003–2012, for the controlling state shareholders, the state owner’s share declined from 47 to 42%; and the non-state controlling owners’ share increased from 33 to 35%. This indicates that although state-controlled firms have a higher concentration of ownership than non-state controlled firms, state ownership has become less dominant. 3.3 Independent variables Political connection. Despite the rising scholarly interest in political connection, there is no ideal measure of political connection. Following Faccio (2006), we adopt a broad definition of political connection: a firm is regarded as politically connected if at least one of its top executives was a senior government official or a member of parliament. We assume that previous substantial work experience in the central government can help SOE managers and local government leaders establish crony arrangements with central leaders and gain greater access to information, both of which can be hugely beneficial in a less transparent political environment. The variable central connection equals 1 if either the chairperson, or general manager, or both, have work experience in the central government agencies or are currently serving as deputies of National People’s Congress (NPC, the legislature) and the Chinese People’s Political Consultative Conference (CPPCC, the political advisory body) at the national level. Note that although SOE top executives are parts of the bureaucratic hierarchy and therefore have government ranks, we assume that those managers who only have work experience in SOEs are not as well-connected as those with central government work experience. The information of central connection is available from 2003 to 2009. How do central connections translate into political favoritism? Previous work experience in the central government can help SOE managers to build channels of influence on policy making in the central government. Access to the central government can be hugely beneficial. For example, Du et al. (2012) find that the evaluation criteria for central SOEs are subject to the influence of political meddling, which gives politically connected firms an edge in receiving good evaluations. Control variables. We expect that a firm’s financial situation will affect its decision of privatization. Thus, we include return on asset (roa), debt to asset ratio (debtratio), log revenue per capita (logrevpc) and log value of total asset (logasset) as control variables. We expect that firms with low profitability, high debt ratio and low productivity are more likely to be controlled by the state. We also include two-digit industry dummies and year dummies to control for industries and time. Controlling for industry fixed effects ensures that the coefficients of interest do not pick up the possibilities that state ownership is explicitly protected in strategic industries, including energy, iron and steel, oil refineries and petrochemicals, communications and heavy machinery, but partial privatization is encouraged in other industries. Controlling for year fixed effects incorporates the possible consequences of legal and regulatory changes in state equity transfers during this period. It is possible that some of the explanatory variables are endogenous to privatization because causality may run the other way. In particular, selling state assets would allow governments to balance the budget and laying off workers would drive up unemployment. In addition, better corporate performance might be the result, not the cause, of privatization. We address these concerns by using lagged control variables. To be sure, lagging these variables provides only a partial solution to the problem because the lagged variables are not strictly exogenous. We will run robustness checks to further address the issue of endogeneity. The summary statistics and definitions of variables are presented in Appendix 1 and 2 available as Supplementary Material. 3.4 Result In models 1 and 2 of Table 1, we estimate probit regressions, where the dependent variable state ownership is a binary indicator. Model 1 only investigates whether firms’ financial situations have any impact on their ownership structure. As expected, large and less productive firms are more likely to be controlled by the state. Model 2 includes the variable political connection. The negative coefficient of political connection indicates that firms are less likely to be controlled by state-equity holders if their managers have connections with the central government. In other words, politically connected firms are more likely to be privatized than politically unconnected firms. It suggests that political connections would give SOE managers greater confidence in transferring state ownership without concerns about losing policy privileges. Table 1. Effect of political and financial factors on privatization, firm-level data 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 Note: Column 1 and 2 report probit regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Columns 3 and 4 report OLS fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 1. Effect of political and financial factors on privatization, firm-level data 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 Note: Column 1 and 2 report probit regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Columns 3 and 4 report OLS fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Models 3 and 4 investigate a subsection of firms in which the controlling owners hold state shares. The dependent variable is state share, defined as the percentage of equities held by the controlling state owner. We use both industry and year fixed effect regressions. The negative coefficient of political connection suggests that state ownership is more dispersed when SOE managers have political connections. It is consistent with the prediction of H1. Politically connected SOEs tend to be less concerned about losing control of corporate governance, and they are more likely to sell state stakes to improve productivity and maximize profits, the key criteria for their performance evaluations. All financial variables have statistically significant effects on state share, but the signs are not always as expected. In the full-set models, SOEs tend to be less productive than non-SOEs. In the sub-set models, SOEs with more concentrated ownership structure are more productive and profitable than those with more dispersed ownership structure. The seemingly inconsistent findings suggest that SOEs in general may be motivated to shift control to private owners with the belief in efficiency gains, whereas more productive SOEs may have less incentive to diversify their ownership structure. The first finding is consistent with the strong cross-national evidence that private firms are in general more productive than SOEs (e.g., Megginson, 2005; Jefferson and Su, 2006; Lardy, 2014), but the second finding is subject to different explanations. One possible explanation is that more productive SOEs tend to be less motivated to dilute their state ownership than unproductive SOEs because dissatisfaction with SOE performance is the most important rationale for privatization (Megginson, 2005). Another possible explanation is that the effect of political connection on state share is conditioned on the productivity of SOEs, which we will examine later. 3.5 Robustness checks There are two possible sources of selection bias. First, political connection may not be a random variable because it may be highly correlated with SOEs’ political status. For example, managers of central SOEs might have closer ties with the central government than their counterparts in local governments. If that is the case, SOEs’ political status might confound the relationship between political connection and privatization outcomes. Second, a firm’s privatization arrangement may not be random (Jefferson and Su, 2006). Politically connected firms may control fewer state shares or have a less concentrated ownership structure to begin with, for any number of reasons, which generates a concern of reverse causality. To address the first selection bias, we create a new variable—firm rank—to capture the political status of firms. It is an ordinal variable with four categories. It takes on value of 3, if a firm’s controlling shareholder is a central SOE.8 A firm is coded as 2, if its controlling shareholder is a SOE supervised by a provincial SASAC.9 A firm is coded as 1, if its controlling shareholder is a SOE supervised by a municipal or county-level government. A firm is coded as 0, if the controlling shareholder is not a SOE. We include firm rank into the baseline model. As shown in Table 2, firm rank is positively associated with state ownership and state share, suggesting that firms’ controlling shareholders are more likely to hold state equities and have more concentrated ownership structure if they are controlled by higher-ranked SOEs. It is because central SOEs are concentrated in strategic industries (i.e. those that are considered economic or political priorities). This finding is consistent with Perotti’s (1995) finding that monopolistic firms in protected industries will tend to be privatized with lower initial sales and possibly a longer time horizon for the shares retained by the government. It is important to note that the effects of central connection remain statistically significant in all models after controlling for firm rank. The substantive effects of central connection actually become greater. Table 2. Firm-level robustness check 1: inclusion of SOE rank 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 Note: Columns 1 and 2 report regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Column 1 uses a probit model and column 2 uses a fixed effect model. Column 3 reports fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 2. Firm-level robustness check 1: inclusion of SOE rank 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 Note: Columns 1 and 2 report regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Column 1 uses a probit model and column 2 uses a fixed effect model. Column 3 reports fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. It is also possible that connections with the central government and national legislature have different effects on privatization. On the one hand, career-focused bureaucrats are motivated to polish their political performance to please the superordinate government. On the other hand, in the absence of competitive elections, seats in the NPC or CPPCC have little real power and carry little institutional incentive for deputies to please voters, but they are the symbols of political recognition for social elites and successful business executives. Therefore, we create two variables to measure potentially different political connections—bureaucratic connection and legislative connection and then conduct a robustness check to identify their individual effects.10 The results, presented in Appendix 3 available as Supplementary Material, indicate that both bureaucratic and legislative connections have negative effects on privatization. To test the possible marginal effect of political connection on privatization conditioned upon firm productivity, we include an interaction variable between central connection and productivity in the specifications of Table 2. The interaction term is not statistically significant, indicating that politically connected SOEs do have more incentive to dilute state ownership, regardless of the firm’s productivity. The regression results are presented in Appendix 4 available as Supplementary Material. To address the concern of reverse causality, we first include a lagged dependent variable as a control. The result is presented in Appendix 5 available as Supplementary Material. Political connection still has significantly negative effect on state ownership but not on state share. This should serve as some evidence that endogeneity is not an immediate concern, but we need to address this concern directly. A more effective method of controlling confounding is matching, which selects observations to ensure the potential confounding variables are evenly distributed in the two groups being compared. The entropy balancing method developed by Hainmueller (2012), by recalibrating the unit weights, can effectively adjust for systematic and random inequalities in representation. We employ this method to create a comparable control group by reweighting the data from the control group to match a set of moments from the data of the treated group. As noted earlier, the share structure reform was implemented between 2005 and 2007, leading to a large-scale conversion of non-tradable shares to tradable shares. Thus, we needed to achieve full balance of firm rank, ownership structure, industry, province and financial variables in 2004. The results of the balancing procedure are shown in Table 3. The treated group had 156 politically connected firms in 2004. On average, these firms have a greater level of state ownership, higher returns and lower debt levels than the controlled group of 1356 unconnected firms. After the balancing, the weighted control group has the same average values across all the relevant covariates. Table 3. Results of entropy balancing Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Note: This table reports results of entropy balancing between connected and unconnected listed companies. The treatment group has 156 units. The unweighted control group has 1356 units. The sum of the control weights equals 156. Table 3. Results of entropy balancing Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Note: This table reports results of entropy balancing between connected and unconnected listed companies. The treatment group has 156 units. The unweighted control group has 1356 units. The sum of the control weights equals 156. With the two groups balanced in the pretreatment period, we can investigate the unbiased effects of political connections on privatization outcomes after the treatment.11 The dependent variables are changes in state ownership and state share during the treatment period. Δstate_ownership is a discrete variable equivalent to −1 if the controlling ownership changes from state equity to nonstate equity, 1 if the controlling ownership changes from non-state to state equity, and 0 if there is no change. We use both ordered probit and OLS models to estimate the effects of prereform political connection on state ownership change. Δstate_share is the change in the equity share owned by the controlling state shareholder. A negative value means decreased equity share and a positive value means increased equity share. All other variables on the left hand side, except for firm rank, use the values for 2004. Table 4 presents the estimates of the balancing model. The negative association between central connection and change in ownership structure means that the controlling owner of a firm is more likely to change his shareholding form from state to non-state equity. Likewise, the negative association between central connection and change in state share suggests that a firm’s controlling state owner is more likely to shed their state equities when he has political connections with the central government. All these results support the hypothesis that SOEs adopt more aggressive privatization schemes when their executives are politically connected. Table 4. Firm-level robustness checks 2: entropy balancing cross-sectional estimates DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 Note: This table reports results of entropy weighted cross-sectional regressions results of change in ownership structure for listed firms after the period of the share structure reform in 2005–2007. The variables for entropy balancing include state ownership_04, ROA_04, debt ratio_04, log revenue per capita_04, log asset_04, industry, province and firm rank. Δ state_ownership is the change in equity type of the controlling shareholder (−1 if the equity type changed from state equity to non-state equity, 1 if changed from non-state equity to state equity, 0 if no change). Δstate_share is the change in equity share owned by the controlling state shareholder (negative value means reduced equity share and positive value means increased equity share). Column 1 uses an ordered probit model; Columns 2 and 3 use an OLS model. Robust standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 4. Firm-level robustness checks 2: entropy balancing cross-sectional estimates DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 Note: This table reports results of entropy weighted cross-sectional regressions results of change in ownership structure for listed firms after the period of the share structure reform in 2005–2007. The variables for entropy balancing include state ownership_04, ROA_04, debt ratio_04, log revenue per capita_04, log asset_04, industry, province and firm rank. Δ state_ownership is the change in equity type of the controlling shareholder (−1 if the equity type changed from state equity to non-state equity, 1 if changed from non-state equity to state equity, 0 if no change). Δstate_share is the change in equity share owned by the controlling state shareholder (negative value means reduced equity share and positive value means increased equity share). Column 1 uses an ordered probit model; Columns 2 and 3 use an OLS model. Robust standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. 4. Provincial-level analysis H2 suggests that provincial leaders with central connections are less likely to privatize SOEs. To test this hypothesis, we aggregate firm-level data into a provincial-level dataset based on firms’ registration addresses. We then investigate the effects of macroeconomic and political factors on the progress of privatization at the provincial level. The dependent variable, SOE share, refers to the percentage of state-controlled listed firms in the total listed firms. For example, in 2003, Hebei province had 31 listed firms, of which 27 were controlled by state-equity holders. In 2012, Hebei had 49 listed firms, of which twelve were controlled by state-equity holders. Thus, SOE share declined from 87% in 2003 to 24% in 2012. The distribution of SOE shares by province is presented in Figure 1. It is clear that the proportion of SOEs in the total listed firms declined in all provinces between 2003 and 2012, but the pace varies considerably across provinces. Figure 1. View largeDownload slide The share of SOEs in listed companies by province. Figure 1. View largeDownload slide The share of SOEs in listed companies by province. To be sure, the decline in SOE shares in the listed firms could be due not just to privatization, but also to the entry of de nova private firms, that is, those with no state-owned predecessor. Because the stock market was mainly used as the means of raising capital for SOEs and provincial governments control the quota of listed firms each year, the number of private firms allowed to list on the stock market should also reflect provincial governments’ preferences on privatization. 4.1 Independent variables Central connection measures whether a provincial leader has previously served as a central party or government official before his provincial post. We code central connection as a binary variable that equals 1 if a provincial leader had substantive central experience before taking provincial positions, and 0 otherwise.12 We consider working as a senior official with the rank of bureau director (juzhang) or above in central government agencies as substantive central experience. We collected the biographical information of provincial party secretaries and governors in each province between 2003 and 2012. Information on these provincial leaders was gathered from zheng tan wang (political forum network) and baidu baike (baidu encyclopedia).13 For example, Li Yuanchao and Liang Baohua served as party secretary and governor of Jiangsu, respectively, from 2002 to 2007. Li Yuanchao was the vice minister of culture (1996–2000) before serving as the party secretary of Jiangsu, so Jiangsu’s central connection (party) is coded as 1 during this period. Liang Baohua did not have previous work experience in the central government, so Jiangsu’s central connection (government) is coded as 0 during this period. 4.2 Control variables We expect that a provincial government’s decision to privatize SOEs may be affected by its economic situation. Financial distress would create incentives for the government to sell state assets to private parties. So our first control variable is balance, measured as fiscal surplus or deficit as a percentage of fiscal revenue ((revenue-expenditure)/revenue).14 A positive number indicates surplus and a negative number indicates deficit. We also expect that unemployment pressure, economic growth and development level affects a province’s privatization policy. High unemployment would make local governments more cautious about promoting privatization. High economic growth would reduce a government’s incentive to sell state equity. Thus, we include GDP per capita (gdppc), GDP growth (growth) and the unemployment rate (unemployment) as control variables. All of the control variables are lagged for one year to mitigate the concern of reverse causality. 4.3 Result In our benchmark regression, presented in column 1 of Table 5, we find that balance is positively associated with SOE ratio, indicating that financial pressure does play an important role in local governments’ privatization decisions. As expected, provinces with large deficits have a lower percentage of listed firms controlled by the state because they have stronger incentives to sell state assets to raise revenues.15 Provinces with higher economic growth rates are less likely to sell state assets, though wealthier provinces are more likely to privatize SOEs. Table 5. Effects of political and economic factors on ownership structure, provincial level DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 5. Effects of political and economic factors on ownership structure, provincial level DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. In the second and third regressions, we include the variable central connection to capture the influence of provincial leaders’ (i.e. party secretaries and governors) ties to the central government. The results show that provincial leaders with central ties have a significant effect on the province’s privatization schemes. Provinces will be less likely to privatize SOEs if their leaders have substantive central experience. It suggests that ties to the central government may change provincial leaders’ priority orders of governance because they tend to put political and social objectives ahead of economic ones. 4.4 Robustness checks Provincial leaders’ SOE experience and tenure. A provincial government’s privatization decision may be affected by its leader’s other experiences. A provincial leader who just started his tenure might be more willing to take risks than one who is about to leave his post. The former might be more likely to use privatization to promote economic growth, whereas the latter might hold off on privatization to maintain social stability. Second, work experience in SOEs could make provincial leaders not just more supportive of SOEs’ political roles, but also more aware of SOEs’ problems. Thus, we code two indicators based on detailed biographical information of provincial leaders. SOE experience measures whether the leader has working experience in SOEs. Tenure measures the number of years a provincial leader served in the position. As shown in columns 1 and 2 in Table 6, provincial leaders’ tenure and SOE experience have no significant effect on the privatization scheme, suggesting that central connection does have an independent effect on provincial governments’ privatization decisions. Table 6. Province-level robustness check DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Column 1 and 2 use the full sample; column 3 and 4 exclude the observations of Beijing and Shanghai. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 6. Province-level robustness check DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Column 1 and 2 use the full sample; column 3 and 4 exclude the observations of Beijing and Shanghai. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Exclusion of outliers. To address the possible concern of a skewed geographic distribution of central SOEs, we dropped the observations for Beijing and Shanghai, two municipalities with more central SOE headquarters than other provinces. Moreover, SOEs located in Beijing and Shanghai may enjoy better information sharing and communications with the central government. As shown in columns 3 and 4 in Table 6, provincial leaders’ central connection still has a significantly positive effect on state stake, although the degree of significance declines slightly, indicating a robust correlation between provincial leaders’ central connections and their privatization schemes. An alternative measure of political connections. A popular measure of political connection is based on factions, which is a distinct characteristic of Chinese politics (Nathan, 1973; Shih, 2008). Factional ties with various top leaders are an important factor that affects local officials’ ranking in the party hierarchy (Shih et al., 2012). According Shih et al. (2016), a provincial governor or party secretary is considered to have factional ties if he/she and a member of politburo standing committee overlapped for one year or over within the same ministerial level work unit. Therefore, the measures of factions are based on less transparent ties between tops leaders and provincial officials in a specific timeframe. The results, presented in Appendix 6 available as Supplementary Material, also modestly support the second hypothesis. A provincial party secretary with factional ties prefers less privatization. Moreover, even when factional ties are controlled, provincial leaders’ central experience is still positively associated with state shares, indicating that connected provincial leaders prefer less privatization regardless of their factional ties with top leaders. Collective effects of political connections. The two hypotheses indicate that political connections at firm and provincial levels have different effects on vested interests’ preferences on privatization, but it is unclear whether these effects reinforce or undermine each other. To test the collective effect of the two variables, we conduct two tests. First, we include a variable leader connection to capture the effect of provincial leaders’ political connection in the firm-level test. We code it as a binary variable that equals 1, if a provincial party secretary or governor had substantive central experience before taking provincial positions, and 0 otherwise. In all models, manager connection consistently has significant negative effect on state ownership and state share, but the effects of leader connection are ambiguous. When the provincial party secretary has close central ties, firms are less likely to be controlled by state owners but tend to have more concentrated ownership structure if they are state owned. When the provincial governor has close central ties, firms are more likely to be controlled by state owners. The coefficients of leader connection and firm connection have different signs in three out of four models, suggesting that political connections have different impacts on firm managers and local leaders with respect to their preferences on privatization. We then include an aggregate variable firm connection, defined as the average degree of central connection of managers in total listed firms in a province, in the provincial-level test. Since the information for firm connection is only available from 2003 to 2009, we have a smaller number of observations. Firm connection has a significantly negative effect on the dynamics of privatization, indicating that firm connection does have an opposite effect on privatization progress. The coefficients on leader connection, while still having the positive sign, are no longer statistically significant. In general, these two tests suggest that political connections have different impacts on the preferences of SOEs and local leaders on privatization, but they do not necessarily reinforce or undermine each other. The results are presented in Appendix 7 and 8 available as Supplementary Material. 5. Conclusion Conventional accounts of the politics of economic development portray the major obstacle to economic reform as powerful vested interests who manipulate politicians to advance their own empires at the expense of the social interest. Whether the economic reforms can succeed or stall will depend on how the government holds off vested interests and generates support for deeper transformation. Our article provides a more nuanced perspective on vested interests for understanding the political dynamics of economic reform. Although economic reforms are often treated as a coherent package that induce political elites to form ‘special interests’ to support or oppose it, attitudes regarding the implementation of reforms are not necessarily consistent across policy areas and/or over time. Rather, different political objectives surrounding each reform should give rise to distinct strategic choices for vested interests. They may support reform in specific areas but block reforms in others, as long as they expect the reform to increase their payoffs. Although both SOE managers and local government leaders are political entrepreneurs under the nomenklura system, their different career prospects create distinct preferences for privatization. Politically connected firms are more likely to pursue rather than block privatization plans. In contrast, local government leaders, constrained by multilateral performance criteria combining economic and political objectives, have more ambiguous attitudes toward privatization. We argue that political connections play an important role in shaping the preferences of SOE managers and provincial leaders regarding privatization in China, albeit in divergent directions. On the one hand, political connections reassure SOE managers that their privileges will be preserved, therefore encouraging them to embrace privatization to maximize their profitability. On the other hand, political connections give provincial leaders more leeway to pursue noneconomic policy goals, therefore reducing their motives to use privatization to generate revenue and boost economic growth. Our multilevel empirical findings provide support to these hypotheses. At the firm level, SOEs are more likely to implement privatization plans when their managers are politically connected. At the provincial level, a province will pursue less aggressive privatization schemes when the provincial leaders have close political ties with the central government. Our results do not indicate that the Chinese experience of partial privatization is a unique model of privatization. Rather, the formation of mixed ownership consortia has increasingly become the global norm because it enables the government to implement privatization programs more easily while preserving some degree of influence in the partially privatized firms (Musaccio and Lazzarini, 2014). This trend may not only raise the conventional question of how the agency conflict between the controlling and the minority shareholders can be reduced as corporate ownership and control can be separated to the benefit of the large shareholders (Porta et al., 1999; Claessens et al., 2000), but also highlight the conflict of interest inherent in the state’s dual role as shareholder and corporate governance regulator, which can have unintended consequences well beyond potential corporate mismanagement (Pargendler, 2012). Partial privatization might be an effective arrangement to reduce resistance to reforms, but, with the potential mismatch of ownership and control, it may be more difficult to establish hard budget constraints for firms. Despite the sweeping and widespread SOE restructuring, SOEs remain the key vehicle for policy-driven investments, which are the major contributor to China’s rapidly rising corporate debt since 2008 (IMF, 2016). Understanding the political logic of the peculiar nature of partial privatization also contributes to the debate about China’s future, which may depend on the ability of the Chinese government to implement the needed reforms (e.g. Shambaugh, 2013, 2016). In recent years, the Chinese government has increasingly relied on the ‘top-level design’ (dingceng sheji) to launch important reforms, aiming to overcome strong resistance of vested interests. But various self-seeking state agencies form contingent interests through the myriads of political and business calculations, leading to inconsistent implementation of reform agendas. With the increasing embeddedness and declining autonomy for policymakers, the once well-performing developmental state models now face serious challenges. In the absence of collective responsibility, politically powerful groups try to manipulate economic reforms for their own purposes rather than for structural transformation. As such, China is falling short of its own objectives for reform. Even a Chinese state think tank has admitted that China’s state capacity for policy implementation has been weakened, indicated by the reform stalemate since 2012 (Buckley, 2017). From a broader perspective, it also helps explain why some countries or economies may be stuck in traps with little or no reform, but also indicates ways to break out of them. It suggests that when designing major economic reforms, particularly concerning privatization, reassuring vested interests is necessary for building a new base of political support for economic reforms. Over the long run, implementing economic reforms requires not only breaking up the old equilibrium by weakening supporters of the status quo, but also consolidating a new equilibrium that institutionalizes the new bases of support that have emerged. Supplementary material Supplementary material is available at Socio-Economic Review Journal online. Acknowledgements The research for this article is partially supported by the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning. I am grateful to Zhang Jianjun and Victor Shih for generously sharing their data. I thank Jorgen Delman, Li Hui, Tang Min, Tang Shiping, Zhu Jiangnan and participants of seminars at Fudan University, University of Hong Kong, University of Copenhagen and “Five-corners-field School” for their helpful comments. I also thank Pippa Morgan for excellent research assistance. Footnotes 1 According to the World Bank, the proceeds generated from 37 large privatization transactions in China accounted for almost one-third of global privatization revenues in 2000–2008 (Nellis, 2012). 2 State shares refer to equities held by governmental agencies or authorized institutions on behalf of the state. Legal-person shares refer to equities owned by companies or institutions with a legal-person status. 3 The general rule is that shareholders owning 5% or more of a company’s outstanding shares can sell their nontradable shares after two years, whereas owners of less than 5% can sell after one year. 4 The performance evaluation guidelines include general and industry-specific targets. General targets included total profit and rate of return on equity, which account for 30 and 40% of the performance score, respectively. The industry-specific targets account for 30% of the performance score (State Council, 2012). 5 For example, Wang Min, who spent his early years as a university professor, had a reputation of promoting rapid privatization while serving as the mayor of Suzhou. He set a record of restructuring 1034 SOEs in one and a half years, which has helped his promotion to the top provincial leader in SOE-heavy provinces of Jilin and Liaoning (Caijing Magazine, 2016). In contrast, Li Keqiang, who was parachuted from the secretary of Chinse Communist Youth League to serve as party secretary of Henan and then Liaoning provinces before moving up to the central government, has not been seen as a particularly strong supporter of privatization in his tenure as a provincial leader. 6 State-invested enterprises (SIEs) consist of four types of enterprises including state-owned enterprises (guoyou duzi qiye), state-owned companies (guoyou duzi gongsi), state-controlled shareholding companies (guoyou ziben konggu gongsi) and state-invested shareholding companies (guoyou ziben cangu gongsi). See Xinhua (2014). 7 State-owned legal-person shares refer to shares directly held by SOEs. 8 The list of SASAC-supervised central SOEs is available at http://www.sasac.gov.cn/n1180/n1226/n2425/. The list of central SOEs supervised by other government agencies was collected from various websites. 9 The list of provincial SOEs was collected from the websites of provincial SASACs. 10 Bureaucratic connection is a dummy variable that equals 1 if either the chairperson of the board, the general manager, or both previously held central government positions; 0 otherwise. Legislative connection is a dummy variable that equals 1 if either the chairperson, the general manager, or both are representatives of NPC or CPPCC; 0 otherwise. 11 This method is similar to Truex’s (2014), who attempts to distinguish the possible confounding influences of formal parliament representation and informal political connections on firms’ financial performance. He finds that becoming a NPC representative will subsequently boost a firm’s profitability, suggesting that formal parliament representation has a direct causal effect on firms’ financial performance, but he also notes that firms that gained NPC representation have better political connections. Political connections are indeed built up and strengthened through executives’ previous work experiences. 12 Since all provincial leaders are deputies of the NPC, it is no longer appropriate to use NPC membership as a measure of political connection. 13 The database of provincial leaders is available at www.st360.cn/jgzyjl/ljjl. 14 The extrabudgetary account was reported separately in the official statistics before 2010, so we calculate balance using different formulas for 2003–2009 and 2010–2012. For the former, balance = ((budgetary revenue + extrabudgetary revenue) – (budgetary expenditure + extrabudgetary expenditure)) / (budgetary revenue + extrabudgetary revenue). For the latter, balance = (budgetary revenue – budgetary expenditure)/ budgetary revenue. 15 The sale of land use rights is another important channel for local governments to raise revenue. Land-related revenue accounted for a third of total local revenue at the provincial level (Lu and Sun, 2013). At the county level, land revenue contributed nearly 80% of extra-budgetary revenue (Kung and Chen, 2013). References Bai C. , Li D. , Tao Z. , Wang Y. ( 2000 ) ‘The multitask theory of State Enterprise Reform’ , Journal of Comparative Economics , 28 , 716 – 738 . Google Scholar CrossRef Search ADS Beijing Youth Daily ( 2014, August 25 ) ‘SOEs Executives Rank Top in Executive Pays of Listed Companies, Even When They Make Huge Losses’, Beijing. Biais B. , Perotti E. ( 2002 ) ‘Machiavellian Privatization’ , American Economic Review , 92 , 240 – 258 . Google Scholar CrossRef Search ADS Bortolotti B. , Milella V. ( 2006 ) Privatization in Western Europe: Stylized Facts, Outcomes, and Open Issues. Nota di Lavoro, Fondazione Eni Enrico Mattei, No. 124. Boycko M. , Shleifer A. , Vishny R. ( 1996 ) ‘A Theory of Privatization’ , The Economic Journal , 106 , 309 – 319 . Google Scholar CrossRef Search ADS Brandt L. , Li H. , Roberts J. ( 2005 ) ‘Banks and Enterprise Privatization in China’ , Journal of Law, Economics and Organization , 21 , 524 – 546 . Google Scholar CrossRef Search ADS Bremmer I. ( 2010 ) The End of the Free Market: Who Wins the War Between States and Corporations? Portfolio. Buckley C. ( 2017, March 27 ) ‘In Rare Move, Chinese Think Tank Criticizes Tepid Pace of Reform’, New York, New York Times. Caijing Magazine ( 2016, April 11 ) ‘The Rise and Fall of Wan Min (Wang Min fuchen)’. Beijing Chan V. ( 2014 ) ‘Sinopec Divests Retail Cheaply: Bad Sign for SOE Reform’, accessed at http://blogs.barrons.com/asiastocks/2014/09/15/sinopec-divests-retail-cheaply-bad-sign-for-soe-reform/ on March 15, 2016. Claessens S. , Djankov S. , Lang L. H. P. ( 2000 ) ‘The Separation of Ownership and Control in East Asian Corporations’, Journal of Financial Economics , 58 , 81 – 112 . Google Scholar CrossRef Search ADS Clarke T.s , Pitelis C. (eds) ( 2005 ) The Political Economy of Privatization . New York , Routledge . Denisova I. , Eller M. , Frye T. , Zhuravskaya E. ( 2009 ) ‘Who Wants to Revise Privatization? The Complementarity of Market Skills and Institutions’ , American Political Science Review , 103 , 284 – 304 Google Scholar CrossRef Search ADS Dinc S. , Gupta N. ( 2011 ) ‘The Decision to Privatize: Finance and Politics’ , Journal of Finance , LXVI , 241 – 270 . Google Scholar CrossRef Search ADS Du F. , Tang G. , Young M. ( 2012 ) ‘Influence Activities and Favoritism in Subjective Performance Evaluation: Evidence from Chinese State-owned enterprises’ , The Accounting Review , 87 , 1555 – 1588 . Google Scholar CrossRef Search ADS Estrin S. , Hanousek J. , Kocenda E. , Svejnar J. ( 2009 ) ‘The Effects of Privatization and Ownership in Transition Economies’ , Journal of Economic Literature , 47 , 1 – 30 . Google Scholar CrossRef Search ADS Evans P. ( 1995 ) Embedded Autonomy: States and Industrial Transformation , Princeton , Princeton University Press . Faccio M. ( 2006 ) ‘Politically Connected Firms’ , American Economic Review , March, 369 – 386 . Faccio M. , Masulis R. , McConnell J. ( 2006 ) ‘Political Connections and Corporate Bailouts’ , Journal of Finance , LXI , 2597 – 2635 . Google Scholar CrossRef Search ADS Fernandez R. , Rodrik D. ( 1991 ) ‘Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty’ , American Economic Review , 81 , 1146 – 1155 . Gan J. , Guo Y. , Xu C. ( 2012 ) What Makes Privatization Work? Evidence from a Large-Scale Nationwide Survey of Chinese Firms, Unpublished Working Paper. Hong Kong University of Science and Technology. Guo G. ( 2007 ) ‘Retrospective Economic Accountability under Authoritarianism: Evidence from China ’, Political Research Quarterly , 60 , 378 – 390 . Google Scholar CrossRef Search ADS Haber S. , Razo A. , Maurer N. ( 2003 ) The Politics of Property Rights: Political Instability, Credible Commitments, and Economic Growth in Mexico, 1876–1929 , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Haggard S. , Kaufmann R. (eds) ( 1992 ) The Politics of Economic Adjustment , Princeton , Princeton University Press . Haggard S. , Webb S. ( 1993 ) ‘What Do We Know About Political Economy of Economic Policy Reform?’ , World Bank Research Observer , 8 , 143 – 168 . Google Scholar CrossRef Search ADS Hainmueller J. ( 2012 ) ‘Entropy Balancing for Causal Effects: A Multivariate Reweighting Method to Produce Balanced Samples in Observational Studies’ , Political Analysis 20 , 25 – 46 . Google Scholar CrossRef Search ADS Hellman J. ( 1998 ) ‘Winners Take All: The Politics of Partial Reform in Postcommunist Transition’ , World Politics , 52 , 203 – 234 . Google Scholar CrossRef Search ADS Hellman J. , Jones G. , Kaufman D. ( 2003 ) ‘Seize the State, Seize the Day: State Capture, Corruption, and Influence in Transition’ , Journal of Comparative Economics , 31 , 751 – 773 . Google Scholar CrossRef Search ADS Hellman J. , Schankerman M. ( 2000 ) ‘Intervention, Corruption and Capture: The Nexus between Enterprises and the State’ , The Economics of Transition , 8 , 545 – 576 . Google Scholar CrossRef Search ADS Holz C. ( 2011 ) ‘The Unbalanced Growth Hypothesis and the Role of the State: The Case of China’s State-owned Enterprises’ , Journal of Development Economics , 96 , 220 – 238 . Google Scholar CrossRef Search ADS Huang Y. ( 2008 ) Capitalism with Chinese Characteristics: Entrepreneurship and the State , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Hung M. , Wong T. J. , Zhang T. ( 2012 ) ‘Political Considerations in the Decision of Chinese SOEs to List in Hong Kong’ , Journal of Accounting and Economics , 53 , 43 – 59 . Google Scholar CrossRef Search ADS International Monetary Fund (IMF) 2016 . Resolving China’s Corporate Debt Problem, IMF Working Paper, WP/16/203. Jefferson G. , Su J. ( 2006 ) ‘Privatization and Restructuring in China: Evidence from Shareholding Ownership, 1995–2001’ , Journal of Comparative Economics , 34 , 146 – 166 . Google Scholar CrossRef Search ADS Jia R. , Kudamatsu M. , Seim D. ( 2015 ) ‘Political Selection in China: The Complementary Roles of Connections and Performance’ , Journal of the European Economic Association , 13 , 631 – 668 . Google Scholar CrossRef Search ADS Johnson C. , Beiman I. ( 2007 ) Balanced Scorecard for State-Owned Enterprises: Driving Performance and Corporate Governance , Manila, Asian Development Bank . Kung J. , Chen T. ( 2013 ) Do Land Revenue Windfalls Reduce the Career Incentives of County Leaders? Evidence from China. Unpublished Working Paper. Hong Kong University of Science and Technology. Landry P. ( 2008 ) Decentralized Authoritarianism in China: The Communist Party’s Control of Local Elites in the Post-Mao Era , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Lardy N. ( 2014 ) Markets over Mao: The Rise of Private Business in China , Washington, DC , The Peterson Institute for International Economics . Lin L. , Milhaupt C. J. ( 2013 ) ‘We Are the (National) Champions: Understanding the Mechanisms of State Capitalism in China’ , Stanford Law Review , 65 , 697 – 760 . Li H. , Meng L. , Wang Q. , Zhou L. ( 2008 ) ‘Political Connections, Financing and Firm Performance: Evidence from Chinese Private Firms’ , Journal of Development Economics , 87 , 283 – 299 . Google Scholar CrossRef Search ADS Liao L. , Liu B. , Wang H. ( 2014 ) ‘China's Secondary Privatization: Perspectives from the Split-Share Structure Reform’ , Journal of Financial Economics , 113 , 500 – 518 . Google Scholar CrossRef Search ADS Lieberthal K. , Oksenberg M. ( 1988 ) Policy making in China , Princeton , Princeton University Press . Liu G. , Sun P. , Woo W. T. ( 2006 ) ‘The Political Economy of Chinese-Style Privatization: Motives and Constraints’ , World Development , 34 , 2016 – 2033 . Google Scholar CrossRef Search ADS Lu Y. , Sun T. ( 2013 ) Local Governments Financing Platforms in China: A Fortune or a Misfortune? IMF Working Paper, WP/13/243, International Monetary Fund. Malesky E. , Taussig M. ( 2008 ) ‘Where is Credit Due? Legal Institutions, Connections, and the Efficiency of Bank Lending in Vietnam’ , Journal of Law, Economics, and Organization , 25 , 535 – 578 . Google Scholar CrossRef Search ADS Megginson W. ( 2005 ) The Financial Economics of Privatization , New York , Oxford University Press . Google Scholar CrossRef Search ADS Mertha A. ( 2009 ) ‘Fragmented Authoritarianism 2.0’ , China Quarterly , 200 , 995 – 1012 . Google Scholar CrossRef Search ADS Milhaupt C. , Zheng W. ( 2016 ) Why Mixed-ownership Reform Cannot Fix China’s State Sector, Paulson Policy Memorandum. Ministry of Finance. ( 2008 ) China Finance Yearbook 2008 , Beijing, China Finance Press . Ministry of Finance. ( 2013 ) China Finance Yearbook 2013 , Beijing, China Finance Press . Moe T. ( 2015 ) ‘Vested Interests and Political Institutions’ , Political Science Quarterly , 130 , 277 – 318 . Google Scholar CrossRef Search ADS Montinola G. , Qian Y. , Weingast B. ( 1995 ) ‘Federalism, Chinese style: The Political Basis for Economic Success in China’ , World Politics , 48 , 50 – 81 . Google Scholar CrossRef Search ADS Murillo M. V. ( 2002 ) ‘Political Bias in Policy Convergence: Privatization Choices in Latin America’ , World Politics , 54 , 462 – 493 . Google Scholar CrossRef Search ADS Musaccio A. , Lazzarini S. ( 2014 ) Reinventing State Capitalism: Leviathan in Business, Brazil and Beyond , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Nathan A. ( 1973 ) ‘A Factionalism Model for CCP Politics’ , The China Quarterly , 53 , 34 – 66 . National Statistical Bureau . National Data, accessed at http://data.stats.gov.cn/index on April 15, 2016. Nellis J. ( 2012 ) The International Experience with Privatization: Its Rapid Rise, Partial Fall and Uncertain Future, SPP Research Paper, 5(3), The School of Public Policy, University of Calgary. Pargendler M. ( 2012 ) ‘State Ownership and Corporate Governance’ , Fordham Law Review , 80 , 2917 – 2973 . Pei M. ( 2006 ) China’s Trapped Transition: The Limits of Developmental Autocracy , Cambridge , Harvard University Press . Google Scholar CrossRef Search ADS Pepinsky T. ( 2009 ) Economic Crisis and the Breakdown of Authoritarian Regimes: India and Malaysia in Comparative Perspective , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Perotti E. ( 1995 ) ‘Credible Privatization’ , American Economic Review , 85 , 847 – 859 . Porta R. , Lopez-de-silanes F. , Shleifer A. ( 1999 ) ‘Corporate Ownership Around the World ’, Journal of Finance , LIV , 471 – 517 . Google Scholar CrossRef Search ADS Przeworski A. ( 1991 ) Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Qian Y. ( 2003 ) ‘How Reform Worked in China’. In Rodrik D. (ed.), In Search of Prosperity: Analytic Narratives on Economic Growth , Princeton , Princeton University Press . Rodrik D. ( 1996 ) ‘Understanding Economic Policy Reform ’, Journal of Economic Literature , 34 , 9 – 41 . Roland G. ( 2002 ) ‘The Political Economy of Transition’ , Journal of Economic Perspectives , 16 , 29 – 50 . Google Scholar CrossRef Search ADS Shleifer A. , Treisman D. ( 2000 ) Without a Map: Political Tactics and Economic Reforms in Russia , Cambridge , MIT Press . State-owned Asset Supervision and Administration Commission (SASAC) ( 2003 ) The Interim Guidelines of Operating Performance Evaluation of Persons in Charge of Central SOEs, accessed at http://www.law-lib.com/law/law_view.asp?id=81865 on March 15, 2016. SASAC ( 2012 ) China’s State-Owned Assets Supervision and Administration Yearbook. Schamis H. ( 1999 ) ‘Distributional Coalitions and the Politics of Economic Reform in Latin America’, World Politics , 51 , 236 – 268 . Google Scholar CrossRef Search ADS Schneider B. R. ( 2004 ) ‘Organizing Interests and Coalitions in the Politics of Market Reform in Latin America’ , World Politics , 56 , 456 – 479 . Google Scholar CrossRef Search ADS Shambaugh D. ( 2013 ) China’s Goes Global: The Partial Power , New York , Oxford University Press . Shambaugh D. ( 2016 ) China’s Future , New York , Polity . Shih V. ( 2008 ) Faction and Finance in China: Elite Conflict and Inflation , New York , Cambridge University Press . Shih V. ( 2007 ) ‘Partial Reform Equilibrium: Chinese Style: Political Incentives and Reform Stagnation in Chinese Financial Policies’ , Comparative Political Studies , 40 , 1238 – 1262 . Google Scholar CrossRef Search ADS Shih V. , Jonghyuk L. , David M. ( 2016 ) Chinese Communist Party Leadership Database , San Diego, CA , UC San Diego . Shih V. , Adolph C. , Liu M. ( 2012 ) ‘Getting Ahead in the Communist Party: Explaining the Advancement of Central Committee Members in China’ , American Political Science Review , 106 , 166 – 187 . Google Scholar CrossRef Search ADS Shleifer A. , Vishny R. ( 1994 ) ‘Politicians and Firms’ , Quarterly Journal of Economics , 109 , 995 – 1025 . Google Scholar CrossRef Search ADS State Council ( 2012 ) ‘Provisional Regulations of Performance Evaluation for Central SOEs’ Leaders’, accessed at http://politics.people.com.cn/n/2013/0201/c70731-20408271.html on March 15, 2016. Truex R. ( 2014 ) ‘The Returns to Office in a ‘Rubber Stamp’ Parliament’ , American Political Science Review , 108 , 235 – 251 . Google Scholar CrossRef Search ADS Walder A. ( 2011 ) ‘From Control to Ownership: China’s Managerial Revolution’ , Management and Organization Review , 7 , 19 – 38 . Google Scholar CrossRef Search ADS Whiting S. ( 2004 ) ‘The Cadre Evaluation System at the Grass Roots: The Paradox of Party Rule’, In Naughton B. , Yang D. (eds), Holding China Together: Diversity and National Integration in the post-Deng Era , New York , Cambridge University Press . World Bank ( 2012 ) China 2030: Building a Modern, Harmonious and Creative Society , Washington, DC , World Bank . Xinhua ( 2013 ) ‘Xinhua Insight: China’s New Reforms Must Breach Vested Interests’, accessed at http://news.xinhuanet.com/english/indepth/2013-11/04/c_132857921.htm on March 15, 2016. Xinhua ( 2014 ) ‘Enterprise State-Owned Asset Law of People’s Republic of China’, accessed at http://news.xinhuanet.com/legal/2008-10/28/content_10270416.htm on March 15, 2016. Xu C. ( 2011 ) ‘The Fundamental Institutions of China’s Reforms and Development’ , Journal of Economic Literature , 49 , 1076 – 1151 . Google Scholar CrossRef Search ADS Yosuf S. , Nabeshima K. , Perkins D. ( 2006 ) Under New Ownership: Privatizing China's State-Owned Enterprises , Palo Alto , Stanford University Press . Zheng Tan Wang . accessed at www.st360.cn/jgzyjl/ljjl, on August 1, 2015. © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Socio-Economic Review Oxford University Press

Privatization with ‘Vested Interests’ in China

Socio-Economic Review , Volume Advance Article – Jul 26, 2017

Loading next page...
 
/lp/ou_press/privatization-with-vested-interests-in-china-Ug0WfDr28b
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
1475-1461
eISSN
1475-147X
D.O.I.
10.1093/ser/mwx023
Publisher site
See Article on Publisher Site

Abstract

Abstract Vested interests have been blamed for resisting the reform of state-owned enterprises (SOEs) in China. Yet, various interest groups have heterogeneous interests in privatization. Using both firm- and provincial-level data, we find that SOE managers and local bureaucrats—two key players of privatization—have contingent, rather than vested, interests in privatization, depending partly on their political connections with the central government. On the one hand, political connections motivate SOE managers to privatize more state ownership while retaining managerial control. On the other hand, central connections discourage provincial leaders from using privatization to boost their economic performance. These results shed light on the conditions under which China implements its economic reforms. With the increasing embeddedness and declining autonomy for policymakers, the once well-performing developmental state models now face serious challenges as politically powerful interest groups can manipulate economic reforms for their own purposes rather than for structural transformation. Success in implementing economic reforms depends on a government’s capacity and skill in generating political support and holding off the opposition (Haggard and Webb, 1993). China’s impressive performance in the early stages of reforms can be attributed to two key institutional arrangements. One is its political institutions that allow the government to have ‘embedded autonomy’ (Evans, 1995), which made it easier for policymakers to make difficult decisions without being hampered by interest groups or the need to form cumbersome political coalitions. The other is economic decentralization, in which local governments are highly motivated to engage in competition for growth-enhancing reforms in pursuit of career advancement (Montinola et al., 1995; Xu, 2011). These two factors are inadequate, however, when we seek to account for the complex dynamics of China’s recent experience of economic reforms. The policy-making process, which was characterized by ‘fragmented authoritarianism’ (Lieberthal and Oksenberg, 1988), has become more fragmented and less authoritarian with the increased influence of vested interests (Pei, 2006; Shih, 2007; Landry, 2008; Mertha, 2009). As the World Bank’s report China 2030 puts it, ‘the close links between the government, big banks, and state enterprises have created vested interests that inhibit reforms and contribute to continued ad hoc state interventions in the economy (World Bank 2012)’. Even China’s official Xinhua News Agency (2013) pointed out that ‘some beneficiaries of reform have started to oppose further changes in the country, becoming “powerful vested interests” that obstruct China’s new reforms’. Privatization is an area of reform that has been deeply decentralized and influenced by local governments with the active involvement of managers of state-owned enterprises (SOEs). Vested interests have been blamed for resisting SOE reform. Yet, the reform has been characterized by rapid privatization of ownership but inconsistent policy goals. On the one hand, since 2000 China has been the global leader in privatization transactions with numerous large-scale mergers, management buyouts and takeovers of small SOEs by private firms.1 On the other hand, concerns about rising state capitalism abound as the Chinese government has exerted more effective control on SOEs and increasingly relied on them to pursue various policy goals (Huang, 2008; Bremmer, 2010). We argue that SOE managers and local government leaders—two key players of privatization—are self-seeking agencies with heterogeneous interests in privatization. SOEs are not entrenched interest groups that consistently oppose privatization, and local government leaders are not reformers that always favor privatization. Whether these two politically powerful groups support or oppose privatization depends partly on their connections with the central government. On the one hand, political connections reassure SOE managers that their benefits and privileges will be preserved, therefore enhancing their incentive to push for privatization in pursuit of performance-based personal rewards. On the other hand, political ties with the central government reduce the pressure from provincial leaders to achieve strong economic growth. It gives them more leeway to put non-economic goals ahead of economic performance goals, therefore reducing their incentive to pursue privatization schemes. We test the effects of political connections on the progress of privatization using both firm- and provincial-level data analyses. At the firm level, we find that centrally connected firms are less likely to be controlled by state-equity holders and have more dispersed ownership structures, indicating that connected SOEs tend to pursue more aggressive privatization schemes. The results are robust after we use the entropy balancing method to mitigate potential selection biases. At the provincial level, we find that where leaders have close ties with the central government, SOEs make up a greater share of listed companies. These results suggest that SOE managers and provincial leaders have contingent, rather than vested, interests in privatization, shaped by their political connections with the central government. This study makes three major contributions. First, it contributes to the literature on the political economy of economic reform. Previous studies suggest that the perceived distributional effects of economic reform will form organized interests for or against economic reforms (Fernandez and Rodrik, 1991; Przeworski, 1991; Haggard and Kaufmann, 1992; Rodrik, 1996; Hellman, 1998; Schamis, 1999). But given the uncertain consequences of economic reform and the government’s commitment problem, collective interests turn out to be more fluid and nuanced than standard interest-based models would predict. This study suggests that political connections, as a partial solution to the government’s commitment problem, are an important factor that shapes the preferences of interest groups concerning economic reforms. Second, it provides a new perspective for understanding business–government relations in authoritarian regimes. Unlike the early developmental state literature, which emphasizes an authoritarian government’s role as a helping hand, and the bureaucratic authoritarianism literature, which focuses on the predatory role of government, this study suggests that a web of business–government ties enhances state influence while constraining state autonomy in economic policy making and implementation. In particular, it implies that the pressure for diminishing the state’s interventive role does not just come from private sector or social groups, but could come from SOEs whose agendas conflict with the state’s original plan. Third, it provides empirical evidence for understanding the consequences of China’s state capitalism. Rather than developing a mutually reinforcing relationship between the state and SOEs, state capitalism seems to create interest groups whose agendas often conflict with the state’s objectives. On the one hand, the state continues to maintain widespread influence despite the gradual process of privatization. On the other hand, economic, rather than political and social, motives have increasingly become the key factor shaping the activities of partially privatized SOEs. The seemingly almighty state capitalism has not developed a successful integration of power and wealth. It is not just because SOEs are less efficient than private firms (Lardy, 2014), but also because the divergence of interests between the state and SOEs may be greater than anticipated. This article proceeds as follows. The next section reviews China’s experience of partial privatization. The third section discusses how political connections may shape the preferences of SOE managers and local politicians concerning privatization and presents two testable hypotheses. The fourth section introduces the datasets and variables. In the fifth section, we present the statistical results from both firm- and provincial-level analyses and run various robustness checks. We conclude in the last section. 1. Road to partial privatization in China Reforms of SOEs in China have a complex history that has been characterized by vague and ambiguous initiatives from the central government and decentralized privatization schemes across regions. SOEs have been making the transition to mixed ownership structures since the late 1980s. In the early 1990s, the central government introduced the split-share structure, allowing SOEs to convert a small portion of state ownership into public ownership while remaining in control of the majority state ownership.2 Through this policy, SOEs can issue roughly equal amounts of state, legal person and public shares. Meanwhile, under the slogan ‘zhuada fangxiao’ (grasp the big, let go of the small), Beijing gave local governments the green light to privatize most of the industries that had little national security or fiscal importance. Eighty percent of the small and medium-sized SOEs were privatized (Johnson and Beiman, 2007). For those still controlled by the state, the restriction on the conversion of non-tradable state and legal-person shares to public ones was eventually lifted in 2005, with a grace period of up to three years.3 Because the privatization campaign was carried out without a clear national legal framework, local governments have a great deal of autonomy in deciding their own paces, and they use a variety of methods to respond to local needs. The primary method is direct sales of state ownership to insiders through management buyouts or to outside private owners; this accounts for about 70% of privatization cases in China. Other approaches include public offerings, joint ventures, leasing and employee shareholding (Gan et al., 2012). Local governments also allowed SOEs to negotiate privatization plans with public shareholders. State shareholders can offer extra shares and generous dividend payouts in exchange for public investors’ approval of the equal trading rights of non-tradable shares (Liao et al., 2014). By the end of 2011, 72% of central SOEs and their subsidiaries had been corporatized and almost all local SOEs had been restructured into shareholding companies with multiple shareholders (SASAC, 2012, p. 31). A new round of SOE reforms was unveiled in November 2013, granting a greater space for local governments and SOEs to find their own best practices to establish mixed ownership structures and adopt a more radical ‘wealth management’ approach to privatization. Large-scale partial privatization led to the rapid decline of SOEs’ role in the country’s economy. Between 2000 and 2012, SOEs’ share of industrial assets declined from 47 to 20%, industrial revenues 29 to 12%, industrial profits 19 to 9% and urban employment 35 to 18% (National Statistical Bureau, 2016). The ownership structure of SOEs has also changed profoundly. Between 1992 and 2012, the total assets of SOEs increased 20 times, but their share of state assets dropped from 49 to 28% (Ministry of Finance, 2008, 2013). Simply focusing on the ownership structure, however, reveals little about the actual degree of autonomy SOE managers have over managerial decision-making and influence the state exerts over the firm. There are significant discrepancies between state ownership and actual control of firms. On the one hand, with the massive delegation of managerial discretion and sales of state ownership shares to entrenched insiders, SOE managers enjoy greater managerial autonomy than one might expect given state ownership (Milhaupt and Zheng, 2016). SOE managers are not just granted a great deal of managerial autonomy, they are also compensated with high-powered incentives. Particularly for managers of SOEs controlled by the central government (hereafter central SOEs), their salaries are explicitly tied to the performance evaluation of their companies.4 They could receive bonuses as high as three times their base salaries if their companies receive an ‘A’ in performance rating (SASAC, 2003). The average compensation for a central SOE manager, for example, was 40% higher than the average level for other top executives in 2013 (Beijing Youth Daily, 2014). On the other hand, under the party–state structure, the state can exert greater influence on partially privatized SOEs than what the ownership structure would suggest, reflecting the distinct characteristics of China’s state capitalism (Lin and Milhaupt, 2013). With the creation of the State-owned Asset Supervision and Administration Commission (SASAC) in 2003, the government has reinforced the elite status of mega-SOEs. In the shadow of party control, SASAC executes the power of selection and compensation of top SOE managers, which ensures the state retains substantial control in key corporate decisions. For partially privatized SOEs, minority private shareholders have no effective right to select executive management teams. Their sole right is to receive dividends at the discretion of state shareholders (Yosuf et al., 2006, p. 90). ‘The SOE reform is now a half-way business’, an investment strategist warned, ‘local governments or central SOEs appear willing to undertake reforms which are beneficial to the vested interests, but seem reluctant to do anything more than that’ (Chan, 2014). 2. Vested interests in privatization Earlier studies on the politics of economic reform tend to portray vested interests as organized groups opposing economic liberalization. The success of economic reform hinges on the ability of the government to mobilize political support and overcome opposition from vested interests (Przeworski, 1991; Haggard and Webb, 1993; Hellman, 1998; Denisova et al., 2009). In the early stages of transition, entrenched winners of initial economic reform, primarily SOEs or former SOEs, are motivated—and will have the power to—block future reform to generate concentrated rents for themselves while imposing high social costs (Hellman, 1998; Shleifer and Treisman, 2000). The power of vested interests lies in their ‘formal ties to the state, legacies of previous state ownership, and more frequent interactions with public officials’. In the later stages of economic transition, however, new private firms may organize to use state capture as a substitute to compete against incumbent firms for political influence (Hellman et al., 2003). This analysis of organized interest groups has been widely applied to studies of economic reforms and political transitions in authoritarian regimes (Pei, 2006; Shih, 2007; Malesky and Taussig, 2008; Pepinsky, 2009). In all these studies, vested interests are considered powerful, and sometimes coherently organized, groups that prefer the status quo to reforms. But studies on business–government relations in Latin America suggest that business interests are more fluid, ambiguous and even conflicted, and are affected by reform packages and firm-level characteristics (Murillo, 2002; Schneider, 2004). Even groups that end up benefiting from the reform may be initially in favor of the status quo if they fear that the reform will make them worse off (Fernandez and Rodrik, 1991; Przeworski, 1991). Given this nuanced portrait of business interests, it is prudent to say that vested interests are not necessarily the enemies of reforms, but under what conditions can vested interests be coopted to support policy changes? The cooptation of vested interests often occurs through the creation of new rents to replace old ones or compensating transfers to the losers from reform (Roland, 2002; Shleifer and Treisman, 2000). Vested interests may become reformers if they expect that disruptive developments would make them worse off, or if they believe that institutional changes will be beneficial and will not involve downside risk (Moe, 2015). In regards to privatization, SOE managers represent a key vested interest group who would oppose privatization as they are concerned about losing their old rents—privileged political statuses and financial subsidies, but their position as anti-privatization stakeholders may change if they expect to receive new rents which may give them additional wealth or political power. The effectiveness of the cooptation strategy, however, is contingent on the government’s ability to make a credible commitment (Roland, 2002). To reduce opposition from SOE managers, the government needs to secure their position as the residual claimant after restructuring SOEs. As long as SOE managers are assured that their policy privileges will continue after partial privatization, they would have little incentive to block the reform. Rather, high-powered incentives can encourage SOE managers to shed more unproductive assets through partial privatization. Thus, partial privatization may become a favorable choice because it could provide SOE managers with a significant ownership stake in the firm and reduced outside government intervention without losing their privileged status (Walder, 2011). Without such commitment from the government, SOE managers will behave like vested interests as expected: they are likely to block the privatization scheme in the first place. However, there is another commitment problem in the process of partial privatization. Perotti (1995) argues that if private investors are concerned about post-privatization government interference, they will be less interested in buying government shares or they will request upfront discounts as compensation for risk. Therefore, if the government could not credibly commit to withdrawing its intervention from corporate governance, SOE reforms are likely to fail (Qian, 2003). These two commitment problems are mutually exclusive, however. If the government makes a non-interference commitment to private investors, it will not be able to guarantee benefits to SOE managers because the government has little control over profit distribution in privatized firms. If the government makes commitment to SOE managers, it will have to intervene in the privatized firms to secure compensating transfers. Thus, the government faces a dilemma in the process of privatization. A credible commitment to SOE managers will violate a commitment to private investors, and vice versa. Because no institutional arrangement could possibly resolve the commitment dilemma faced by the government, crony arrangements may become a partial solution, particularly in authoritarian regimes where the government has inherently weak commitment power in the absence of institutional constraints (Haber et al., 2003). Having patrons or representatives in the government not only helps an interest group influence general policy making, but also allows the government to guarantee that the property rights of a subset of asset holders will be protected. Recent studies show that political connections enable powerful firms to gain greater access to particularistic benefits and boost their market values (Hellman and Schankerman, 2000; Faccio, 2006; Faccio et al., 2006; Li et al., 2008; Malesky and Taussig, 2008). Connected SOE managers also receive more personal benefits than those in unconnected firms (Hung et al., 2012). A politically connected SOE manager will be more likely to accept policy changes, if political connections enhance the credibility of the government’s commitment to post-privatization compensating transfer. They may even support privatization because it is in their interest to obtain more management autonomy and ‘depoliticize’ the firm. This discussion yields our first hypothesis. H1: Politically connected SOE managers pursue more aggressive privatization schemes. While SOE managers appear to be vested interests with clear-cut preferences in opposing privatization, local government leaders’ interest in privatization is more ambiguous. On the one hand, controlling SOEs gives local government leaders freedom to use SOEs as critical policy tools to channel public investments and maintain employment, but these benefits depend on the interaction between governments, financial institutions and enterprise managers (Brandt et al., 2005). An increase in market competitiveness, on the other hand, reduces the advantages of state ownership. When the SOE’s budget becomes harder, local leaders may find controlling state ownership less valuable compared to the benefits received from selling the firm. They may be motivated to think SOEs should be privatized if they expect that privatization can increase tax revenue and enhance enterprise efficiency (Boycko et al., 1996; Liu et al., 2006). Although private firms tend to be more efficient and productive than SOEs, the effect of privatization on firm productivity and economic growth is inconclusive in developing countries and transition economies (Clarke and Pitelis, 2005; Megginson, 2005; Bortolotti and Milella, 2006; Estrin et al., 2009). Given the uncertain economic benefits of privatization, whether and how to privatize are essentially the result of political calculations. It is most likely to occur when a local politician cannot obtain large benefits from public firms any more (Shleifer and Vishny, 1994). In democracies, politicians’ motives for privatization depend on their ability to extract political benefits and voting support from privatization (Biais and Perotti, 2002; Dinc and Gupta, 2011). In authoritarian regimes, local politicians also need to survive political competition, but they are, first and foremost, subject to upward rather than downward accountability. In other words, local politicians’ primary motive is to please their superiors rather than their constituencies. In China, where the political regime is characterized by ‘regionally decentralized authoritarianism’ that combines centralized personnel control and a regionally decentralized economic system (Xu, 2011), career-oriented local leaders are motivated to demonstrate political loyalty to gain political credits and deliver strong economic performance to gain economic credits. Jia et al. (2015) find that political connections increase the likelihood of promotion for local leaders, but political connections alone cannot guarantee local leaders’ promotion unless they can deliver strong economic performance. While connections and performance are complements in the promotion of provincial leaders, unconnected provincial leaders will have to rely more on delivering strong economic performance to increase their odds of promotion. The relative importance of political credits and economic credits vary at different levels of government. For leaders at lower-level (county or township) governments, economic performance is a strong predictor of their promotion (Whiting, 2004; Guo 2007; Landry 2008). For provincial government leaders, political factions or connections play a more important role (Shih et al., 2012). Similarly, Bai et al. (2000) argue that higher-level bureaucrats tend to be more concerned about the cost of privatization, such as layoffs of surplus workers and the underprovision of social welfare, whereas lower-level bureaucrats are more likely to pursue the benefits of privatization, such as enhanced profitability and reduced fiscal burdens. Thus, local leaders take both political and economic credits into consideration when implementing privatization plans. They need to find a policy more beneficial for their career prospects than simply pursuing SOE profit maximization (Holz, 2011). Their preferences on privatization may depend on how close they are with the central government.5 We expect that provincial leaders with close ties to the central government will be less likely to pursue privatization because they tend to put non-economic goals (e.g., job creation and social welfare provision) ahead of economic goals whereas those without close ties are more likely to pursue privatization to promote local economic growth. Thus, we have the second hypothesis: H2: Provincial leaders with close ties with the central government pursue less aggressive privatization schemes. 3. Data and empirical strategy To investigate the impact of vested interests and local officials on privatization, we use two datasets. We first use a firm-level dataset to estimate the effect of firm executives’ political connections on firms’ privatization outcomes. Then we use an aggregated dataset to estimate the effect of local leaders’ political connections on the progress of privatization at the provincial level. 3.1 Firm-level analysis The firm-level dataset includes firms that were publicly listed on the main boards of the Shanghai and Shenzhen Stock Exchanges between 2003 and 2009 (ranging from 1285 to 2457). It was assembled from three sources: the China Stock Market and Accounting Research (CSMAR) database for accounting and financial information, the Sinofin database compiled by the China Center for Economic Research (CCER) for information on ownership structure and major shareholders and the annual reports of listed firms for biographical information of firms’ top executives (chairpersons and general managers). In instances where biographical information was not detailed enough in the annual reports, we performed an Internet search to collect company executives’ career information (e.g. bureaucratic or legislative experiences). In addition, all financial variables have been winsorized at the 0.1% and 99.9% levels to minimize the impact of outliers. The panel is unbalanced due to the timing of listing, mergers and acquisitions and missing information for some financial variables. 3.2 Dependent variable Our primary interest is SOEs, but the definition of SOE has evolved over time. Traditional SOEs only refer to enterprises directly owned by the state or state agencies, but the Chinese government has broadened the scope of SOEs to state-invested enterprises (guojia chuzi qiye), which includes those indirectly controlled by the state.6 Since our database does not have information on indirect shareholding, we focus on the changes in types and structures of direct controlling shareholders. There are three major types of shareholder: state, legal person and public shareholders. State shareholders hold state shares (guojia gu) or state-owned legal-person shares (guoyou faren gu).7 Legal-person shareholders hold all other types of non-tradable shares (faren gu). Public shareholders hold all types of publicly tradable shares (gongzhong gu). State ownership. Our first dependent variable is a binary variable that equals 1 if the controlling shareholder holds state equities and 0 otherwise. In 2003, 68% of controlling shareholders of listed firms held state equities. In 2012, the trend changed with only 14% of controlling shareholders holding state equities. State share. To be sure, classifying firms based on whether the majority or dominant owner is state cannot neatly separate firms into state and private domains as many firms have always had hybrid forms of ownership in which state units are sometimes the underlying owner (Lardy, 2014). Therefore, a change in shareholding forms from state equity to non-state equity does not necessarily mean that the state has relinquished control. It could simply be the result of regulatory change that allows state owners to convert their non-tradable state equities into tradable public equities. Thus, we use another variable, state share, which is defined as the proportion of equity owned by the controlling state shareholder, to capture changes in the degree of ownership concentration. During the period 2003–2012, for the controlling state shareholders, the state owner’s share declined from 47 to 42%; and the non-state controlling owners’ share increased from 33 to 35%. This indicates that although state-controlled firms have a higher concentration of ownership than non-state controlled firms, state ownership has become less dominant. 3.3 Independent variables Political connection. Despite the rising scholarly interest in political connection, there is no ideal measure of political connection. Following Faccio (2006), we adopt a broad definition of political connection: a firm is regarded as politically connected if at least one of its top executives was a senior government official or a member of parliament. We assume that previous substantial work experience in the central government can help SOE managers and local government leaders establish crony arrangements with central leaders and gain greater access to information, both of which can be hugely beneficial in a less transparent political environment. The variable central connection equals 1 if either the chairperson, or general manager, or both, have work experience in the central government agencies or are currently serving as deputies of National People’s Congress (NPC, the legislature) and the Chinese People’s Political Consultative Conference (CPPCC, the political advisory body) at the national level. Note that although SOE top executives are parts of the bureaucratic hierarchy and therefore have government ranks, we assume that those managers who only have work experience in SOEs are not as well-connected as those with central government work experience. The information of central connection is available from 2003 to 2009. How do central connections translate into political favoritism? Previous work experience in the central government can help SOE managers to build channels of influence on policy making in the central government. Access to the central government can be hugely beneficial. For example, Du et al. (2012) find that the evaluation criteria for central SOEs are subject to the influence of political meddling, which gives politically connected firms an edge in receiving good evaluations. Control variables. We expect that a firm’s financial situation will affect its decision of privatization. Thus, we include return on asset (roa), debt to asset ratio (debtratio), log revenue per capita (logrevpc) and log value of total asset (logasset) as control variables. We expect that firms with low profitability, high debt ratio and low productivity are more likely to be controlled by the state. We also include two-digit industry dummies and year dummies to control for industries and time. Controlling for industry fixed effects ensures that the coefficients of interest do not pick up the possibilities that state ownership is explicitly protected in strategic industries, including energy, iron and steel, oil refineries and petrochemicals, communications and heavy machinery, but partial privatization is encouraged in other industries. Controlling for year fixed effects incorporates the possible consequences of legal and regulatory changes in state equity transfers during this period. It is possible that some of the explanatory variables are endogenous to privatization because causality may run the other way. In particular, selling state assets would allow governments to balance the budget and laying off workers would drive up unemployment. In addition, better corporate performance might be the result, not the cause, of privatization. We address these concerns by using lagged control variables. To be sure, lagging these variables provides only a partial solution to the problem because the lagged variables are not strictly exogenous. We will run robustness checks to further address the issue of endogeneity. The summary statistics and definitions of variables are presented in Appendix 1 and 2 available as Supplementary Material. 3.4 Result In models 1 and 2 of Table 1, we estimate probit regressions, where the dependent variable state ownership is a binary indicator. Model 1 only investigates whether firms’ financial situations have any impact on their ownership structure. As expected, large and less productive firms are more likely to be controlled by the state. Model 2 includes the variable political connection. The negative coefficient of political connection indicates that firms are less likely to be controlled by state-equity holders if their managers have connections with the central government. In other words, politically connected firms are more likely to be privatized than politically unconnected firms. It suggests that political connections would give SOE managers greater confidence in transferring state ownership without concerns about losing policy privileges. Table 1. Effect of political and financial factors on privatization, firm-level data 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 Note: Column 1 and 2 report probit regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Columns 3 and 4 report OLS fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 1. Effect of political and financial factors on privatization, firm-level data 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 1 2 3 4 DV State ownership State ownership State share State share ROA (lag) 0.093 (0.079) 0.100 (0.079) 5.247*** (1.748) 5.346*** (1.748) Debt ratio (lag) 0.004 (0.014) 0.004 (0.014) 0.520** (0.237) 0.526** (0.237) Log revenue per capita (lag) −0.036*** (0.013) −0.036*** (0.013) 0.932*** (0.205) 0.933*** (0.205) Log asset (lag) 0.281*** (0.016) 0.289*** (0.016) 1.607*** (0.224) 1.649*** (0.225) Central Connection −0.200*** (0.045) −1.207* (0.656) Constant −4.640*** (0.347) −4.640*** (0.347) −8.632* (4.831) −8.632* (4.831) Industry dummy Yes Yes Yes Yes Year dummy Yes Yes Yes Yes N 9497 9497 5414 5414 R2 0.217 0.217 Note: Column 1 and 2 report probit regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Columns 3 and 4 report OLS fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Models 3 and 4 investigate a subsection of firms in which the controlling owners hold state shares. The dependent variable is state share, defined as the percentage of equities held by the controlling state owner. We use both industry and year fixed effect regressions. The negative coefficient of political connection suggests that state ownership is more dispersed when SOE managers have political connections. It is consistent with the prediction of H1. Politically connected SOEs tend to be less concerned about losing control of corporate governance, and they are more likely to sell state stakes to improve productivity and maximize profits, the key criteria for their performance evaluations. All financial variables have statistically significant effects on state share, but the signs are not always as expected. In the full-set models, SOEs tend to be less productive than non-SOEs. In the sub-set models, SOEs with more concentrated ownership structure are more productive and profitable than those with more dispersed ownership structure. The seemingly inconsistent findings suggest that SOEs in general may be motivated to shift control to private owners with the belief in efficiency gains, whereas more productive SOEs may have less incentive to diversify their ownership structure. The first finding is consistent with the strong cross-national evidence that private firms are in general more productive than SOEs (e.g., Megginson, 2005; Jefferson and Su, 2006; Lardy, 2014), but the second finding is subject to different explanations. One possible explanation is that more productive SOEs tend to be less motivated to dilute their state ownership than unproductive SOEs because dissatisfaction with SOE performance is the most important rationale for privatization (Megginson, 2005). Another possible explanation is that the effect of political connection on state share is conditioned on the productivity of SOEs, which we will examine later. 3.5 Robustness checks There are two possible sources of selection bias. First, political connection may not be a random variable because it may be highly correlated with SOEs’ political status. For example, managers of central SOEs might have closer ties with the central government than their counterparts in local governments. If that is the case, SOEs’ political status might confound the relationship between political connection and privatization outcomes. Second, a firm’s privatization arrangement may not be random (Jefferson and Su, 2006). Politically connected firms may control fewer state shares or have a less concentrated ownership structure to begin with, for any number of reasons, which generates a concern of reverse causality. To address the first selection bias, we create a new variable—firm rank—to capture the political status of firms. It is an ordinal variable with four categories. It takes on value of 3, if a firm’s controlling shareholder is a central SOE.8 A firm is coded as 2, if its controlling shareholder is a SOE supervised by a provincial SASAC.9 A firm is coded as 1, if its controlling shareholder is a SOE supervised by a municipal or county-level government. A firm is coded as 0, if the controlling shareholder is not a SOE. We include firm rank into the baseline model. As shown in Table 2, firm rank is positively associated with state ownership and state share, suggesting that firms’ controlling shareholders are more likely to hold state equities and have more concentrated ownership structure if they are controlled by higher-ranked SOEs. It is because central SOEs are concentrated in strategic industries (i.e. those that are considered economic or political priorities). This finding is consistent with Perotti’s (1995) finding that monopolistic firms in protected industries will tend to be privatized with lower initial sales and possibly a longer time horizon for the shares retained by the government. It is important to note that the effects of central connection remain statistically significant in all models after controlling for firm rank. The substantive effects of central connection actually become greater. Table 2. Firm-level robustness check 1: inclusion of SOE rank 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 Note: Columns 1 and 2 report regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Column 1 uses a probit model and column 2 uses a fixed effect model. Column 3 reports fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 2. Firm-level robustness check 1: inclusion of SOE rank 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 1 2 3 State ownership State ownership State share Model Probit Fixed effect Fixed effect ROA (lag) 0.117 (0.089) 0.027 (0.019) 5.150***(1.745) Debt ratio (lag) 0.013 (0.014) 0.004 (0.003) 0.557**(0.237) Log revenue per capita (lag) −0.087***(0.015) −0.021***(0.004) 0.852***(0.205) Log asset (lag) 0.138***(0.018) 0.032***(0.004) 1.540***(0.225) Central Connection −0.319***(0.052) −0.084***(0.013) −1.406**(0.656) Firm rank 0.862***(0.018) 0.251***(0.004) 1.207***(0.254) Constant −1.706***(0.400) −0.042 (0.089) −7.456 (4.828) Year dummies Y Y Y Industry dummies Y Y Y N 9497 9595 5414 R2 0.430 0.220 Note: Columns 1 and 2 report regression results of the effect of political and financial factors on controlling shareholders’ ownership form from 2003 to 2009. The dependent variable state ownership is a binary variable that equals 1 if the controlling shareholder holds state equities or 0 if the controlling shareholder holds non-state equities. Two-digit industry dummies and year dummies are included but are not reported. Column 1 uses a probit model and column 2 uses a fixed effect model. Column 3 reports fixed effects regression results of the effect of political and financial factors on controlling state shareholders’ equity share from 2003 to 2009. The dependent variable state share is the percentage of equity owned by the controlling state shareholder. Two-digit industry dummies and year dummies are included but are not reported. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. It is also possible that connections with the central government and national legislature have different effects on privatization. On the one hand, career-focused bureaucrats are motivated to polish their political performance to please the superordinate government. On the other hand, in the absence of competitive elections, seats in the NPC or CPPCC have little real power and carry little institutional incentive for deputies to please voters, but they are the symbols of political recognition for social elites and successful business executives. Therefore, we create two variables to measure potentially different political connections—bureaucratic connection and legislative connection and then conduct a robustness check to identify their individual effects.10 The results, presented in Appendix 3 available as Supplementary Material, indicate that both bureaucratic and legislative connections have negative effects on privatization. To test the possible marginal effect of political connection on privatization conditioned upon firm productivity, we include an interaction variable between central connection and productivity in the specifications of Table 2. The interaction term is not statistically significant, indicating that politically connected SOEs do have more incentive to dilute state ownership, regardless of the firm’s productivity. The regression results are presented in Appendix 4 available as Supplementary Material. To address the concern of reverse causality, we first include a lagged dependent variable as a control. The result is presented in Appendix 5 available as Supplementary Material. Political connection still has significantly negative effect on state ownership but not on state share. This should serve as some evidence that endogeneity is not an immediate concern, but we need to address this concern directly. A more effective method of controlling confounding is matching, which selects observations to ensure the potential confounding variables are evenly distributed in the two groups being compared. The entropy balancing method developed by Hainmueller (2012), by recalibrating the unit weights, can effectively adjust for systematic and random inequalities in representation. We employ this method to create a comparable control group by reweighting the data from the control group to match a set of moments from the data of the treated group. As noted earlier, the share structure reform was implemented between 2005 and 2007, leading to a large-scale conversion of non-tradable shares to tradable shares. Thus, we needed to achieve full balance of firm rank, ownership structure, industry, province and financial variables in 2004. The results of the balancing procedure are shown in Table 3. The treated group had 156 politically connected firms in 2004. On average, these firms have a greater level of state ownership, higher returns and lower debt levels than the controlled group of 1356 unconnected firms. After the balancing, the weighted control group has the same average values across all the relevant covariates. Table 3. Results of entropy balancing Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Note: This table reports results of entropy balancing between connected and unconnected listed companies. The treatment group has 156 units. The unweighted control group has 1356 units. The sum of the control weights equals 156. Table 3. Results of entropy balancing Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Treat Control Control (weighted) mean Variance mean variance Mean variance Firm rank 1.462 1.502 1.126 1.201 1.462 1.287 State ownership_04 0.628 0.235 0.563 0.246 0.628 0.234 ROA_04 0.039 0.004 0.011 0.0409 0.039 0.004 Debt ratio_04 0.470 0.032 0.531 0.325 0.470 0.041 Log revenue per capita_04 13.32 1.098 13.27 1.496 13.32 1.347 Log asset_04 21.53 1.462 21.21 1.048 21.53 1.308 Industry code 46.65 687.8 49.97 683.4 46.65 634.9 Province code 21.65 59.6 22.88 55.1 21.65 56.3 Note: This table reports results of entropy balancing between connected and unconnected listed companies. The treatment group has 156 units. The unweighted control group has 1356 units. The sum of the control weights equals 156. With the two groups balanced in the pretreatment period, we can investigate the unbiased effects of political connections on privatization outcomes after the treatment.11 The dependent variables are changes in state ownership and state share during the treatment period. Δstate_ownership is a discrete variable equivalent to −1 if the controlling ownership changes from state equity to nonstate equity, 1 if the controlling ownership changes from non-state to state equity, and 0 if there is no change. We use both ordered probit and OLS models to estimate the effects of prereform political connection on state ownership change. Δstate_share is the change in the equity share owned by the controlling state shareholder. A negative value means decreased equity share and a positive value means increased equity share. All other variables on the left hand side, except for firm rank, use the values for 2004. Table 4 presents the estimates of the balancing model. The negative association between central connection and change in ownership structure means that the controlling owner of a firm is more likely to change his shareholding form from state to non-state equity. Likewise, the negative association between central connection and change in state share suggests that a firm’s controlling state owner is more likely to shed their state equities when he has political connections with the central government. All these results support the hypothesis that SOEs adopt more aggressive privatization schemes when their executives are politically connected. Table 4. Firm-level robustness checks 2: entropy balancing cross-sectional estimates DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 Note: This table reports results of entropy weighted cross-sectional regressions results of change in ownership structure for listed firms after the period of the share structure reform in 2005–2007. The variables for entropy balancing include state ownership_04, ROA_04, debt ratio_04, log revenue per capita_04, log asset_04, industry, province and firm rank. Δ state_ownership is the change in equity type of the controlling shareholder (−1 if the equity type changed from state equity to non-state equity, 1 if changed from non-state equity to state equity, 0 if no change). Δstate_share is the change in equity share owned by the controlling state shareholder (negative value means reduced equity share and positive value means increased equity share). Column 1 uses an ordered probit model; Columns 2 and 3 use an OLS model. Robust standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 4. Firm-level robustness checks 2: entropy balancing cross-sectional estimates DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 DV ΔState ownership ΔState ownership ΔState share Model Ordered Probit OLS OLS Central connection_04 −0.327*** (0.118) −0.087*** (0.031) −2.815** (1.142) N 1511 1511 860 R2 0.074 0.025 Note: This table reports results of entropy weighted cross-sectional regressions results of change in ownership structure for listed firms after the period of the share structure reform in 2005–2007. The variables for entropy balancing include state ownership_04, ROA_04, debt ratio_04, log revenue per capita_04, log asset_04, industry, province and firm rank. Δ state_ownership is the change in equity type of the controlling shareholder (−1 if the equity type changed from state equity to non-state equity, 1 if changed from non-state equity to state equity, 0 if no change). Δstate_share is the change in equity share owned by the controlling state shareholder (negative value means reduced equity share and positive value means increased equity share). Column 1 uses an ordered probit model; Columns 2 and 3 use an OLS model. Robust standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. 4. Provincial-level analysis H2 suggests that provincial leaders with central connections are less likely to privatize SOEs. To test this hypothesis, we aggregate firm-level data into a provincial-level dataset based on firms’ registration addresses. We then investigate the effects of macroeconomic and political factors on the progress of privatization at the provincial level. The dependent variable, SOE share, refers to the percentage of state-controlled listed firms in the total listed firms. For example, in 2003, Hebei province had 31 listed firms, of which 27 were controlled by state-equity holders. In 2012, Hebei had 49 listed firms, of which twelve were controlled by state-equity holders. Thus, SOE share declined from 87% in 2003 to 24% in 2012. The distribution of SOE shares by province is presented in Figure 1. It is clear that the proportion of SOEs in the total listed firms declined in all provinces between 2003 and 2012, but the pace varies considerably across provinces. Figure 1. View largeDownload slide The share of SOEs in listed companies by province. Figure 1. View largeDownload slide The share of SOEs in listed companies by province. To be sure, the decline in SOE shares in the listed firms could be due not just to privatization, but also to the entry of de nova private firms, that is, those with no state-owned predecessor. Because the stock market was mainly used as the means of raising capital for SOEs and provincial governments control the quota of listed firms each year, the number of private firms allowed to list on the stock market should also reflect provincial governments’ preferences on privatization. 4.1 Independent variables Central connection measures whether a provincial leader has previously served as a central party or government official before his provincial post. We code central connection as a binary variable that equals 1 if a provincial leader had substantive central experience before taking provincial positions, and 0 otherwise.12 We consider working as a senior official with the rank of bureau director (juzhang) or above in central government agencies as substantive central experience. We collected the biographical information of provincial party secretaries and governors in each province between 2003 and 2012. Information on these provincial leaders was gathered from zheng tan wang (political forum network) and baidu baike (baidu encyclopedia).13 For example, Li Yuanchao and Liang Baohua served as party secretary and governor of Jiangsu, respectively, from 2002 to 2007. Li Yuanchao was the vice minister of culture (1996–2000) before serving as the party secretary of Jiangsu, so Jiangsu’s central connection (party) is coded as 1 during this period. Liang Baohua did not have previous work experience in the central government, so Jiangsu’s central connection (government) is coded as 0 during this period. 4.2 Control variables We expect that a provincial government’s decision to privatize SOEs may be affected by its economic situation. Financial distress would create incentives for the government to sell state assets to private parties. So our first control variable is balance, measured as fiscal surplus or deficit as a percentage of fiscal revenue ((revenue-expenditure)/revenue).14 A positive number indicates surplus and a negative number indicates deficit. We also expect that unemployment pressure, economic growth and development level affects a province’s privatization policy. High unemployment would make local governments more cautious about promoting privatization. High economic growth would reduce a government’s incentive to sell state equity. Thus, we include GDP per capita (gdppc), GDP growth (growth) and the unemployment rate (unemployment) as control variables. All of the control variables are lagged for one year to mitigate the concern of reverse causality. 4.3 Result In our benchmark regression, presented in column 1 of Table 5, we find that balance is positively associated with SOE ratio, indicating that financial pressure does play an important role in local governments’ privatization decisions. As expected, provinces with large deficits have a lower percentage of listed firms controlled by the state because they have stronger incentives to sell state assets to raise revenues.15 Provinces with higher economic growth rates are less likely to sell state assets, though wealthier provinces are more likely to privatize SOEs. Table 5. Effects of political and economic factors on ownership structure, provincial level DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 5. Effects of political and economic factors on ownership structure, provincial level DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 DV SOE share 1 2 3 Balance (lag) 0.052*** (0.020) 0.050*** (0.020) 0.048** (0.020) Unemployment (lag) −0.001 (0.002) −0.001 (0.002) −0.003* (0.002) Growth (lag) 0.023*** (0.003) 0.023*** (0.003) 0.025*** (0.003) Log GDPPC (lag) −0.423*** (0.014) −0.431*** (0.015) −0.426*** (0.014) Central connection (S) 0.038** (0.017) Central connection (G) 0.052*** (0.016) Constant 2.100*** (0.375) 2.161*** (0.374) 1.976*** (0.371) N 307 306 306 R2 0.848 0.850 0.850 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. In the second and third regressions, we include the variable central connection to capture the influence of provincial leaders’ (i.e. party secretaries and governors) ties to the central government. The results show that provincial leaders with central ties have a significant effect on the province’s privatization schemes. Provinces will be less likely to privatize SOEs if their leaders have substantive central experience. It suggests that ties to the central government may change provincial leaders’ priority orders of governance because they tend to put political and social objectives ahead of economic ones. 4.4 Robustness checks Provincial leaders’ SOE experience and tenure. A provincial government’s privatization decision may be affected by its leader’s other experiences. A provincial leader who just started his tenure might be more willing to take risks than one who is about to leave his post. The former might be more likely to use privatization to promote economic growth, whereas the latter might hold off on privatization to maintain social stability. Second, work experience in SOEs could make provincial leaders not just more supportive of SOEs’ political roles, but also more aware of SOEs’ problems. Thus, we code two indicators based on detailed biographical information of provincial leaders. SOE experience measures whether the leader has working experience in SOEs. Tenure measures the number of years a provincial leader served in the position. As shown in columns 1 and 2 in Table 6, provincial leaders’ tenure and SOE experience have no significant effect on the privatization scheme, suggesting that central connection does have an independent effect on provincial governments’ privatization decisions. Table 6. Province-level robustness check DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Column 1 and 2 use the full sample; column 3 and 4 exclude the observations of Beijing and Shanghai. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Table 6. Province-level robustness check DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 DV SOE share 1 2 3 4 Balance (lag) 0.050** (0.020) 0.048** (0.020) 0.054*** (0.020) 0.051** (0.020) Unemployment (lag) −0.002 (0.002) −0.003* (0.002) −0.002 (0.002) −0.003** (0.002) Growth (lag) 0.023*** (0.003) 0.025*** (0.003) 0.019*** (0.004) 0.021*** (0.004) Log GDPPC (lag) −0.431*** (0.015) −0.431*** (0.015) −0.419*** (0.018) −0.416*** (0.017) Central connection (S) 0.043** (0.018) 0.036* (0.017) SOE experience (S) 0.018 (0.017) Tenure (S) 0.001 (0.003) Central connection (G) 0.059*** (0.017) 0.047*** (0.017) SOE experience (G) −0.028* (0.016) Tenure (G) 0.000 (0.003) Constant 2.166*** (0.380) 2.051*** (0.373) 2.522*** (0.389) 2.313*** (0.389) N 306 306 286 286 R2 0.851 0.855 0.853 0.855 Note: This table reports province fixed effects regression results of the effect of economic and political factors on a province’s privatization effort from 2003 to 2012. The dependent variable SOE share is the proportion of firms controlled by state-equity holders in total listed firms. Column 1 and 2 use the full sample; column 3 and 4 exclude the observations of Beijing and Shanghai. Standard errors are in parentheses. Asterisks denote the level of statistical significance: *** 1% ** 5% and * 10%. Exclusion of outliers. To address the possible concern of a skewed geographic distribution of central SOEs, we dropped the observations for Beijing and Shanghai, two municipalities with more central SOE headquarters than other provinces. Moreover, SOEs located in Beijing and Shanghai may enjoy better information sharing and communications with the central government. As shown in columns 3 and 4 in Table 6, provincial leaders’ central connection still has a significantly positive effect on state stake, although the degree of significance declines slightly, indicating a robust correlation between provincial leaders’ central connections and their privatization schemes. An alternative measure of political connections. A popular measure of political connection is based on factions, which is a distinct characteristic of Chinese politics (Nathan, 1973; Shih, 2008). Factional ties with various top leaders are an important factor that affects local officials’ ranking in the party hierarchy (Shih et al., 2012). According Shih et al. (2016), a provincial governor or party secretary is considered to have factional ties if he/she and a member of politburo standing committee overlapped for one year or over within the same ministerial level work unit. Therefore, the measures of factions are based on less transparent ties between tops leaders and provincial officials in a specific timeframe. The results, presented in Appendix 6 available as Supplementary Material, also modestly support the second hypothesis. A provincial party secretary with factional ties prefers less privatization. Moreover, even when factional ties are controlled, provincial leaders’ central experience is still positively associated with state shares, indicating that connected provincial leaders prefer less privatization regardless of their factional ties with top leaders. Collective effects of political connections. The two hypotheses indicate that political connections at firm and provincial levels have different effects on vested interests’ preferences on privatization, but it is unclear whether these effects reinforce or undermine each other. To test the collective effect of the two variables, we conduct two tests. First, we include a variable leader connection to capture the effect of provincial leaders’ political connection in the firm-level test. We code it as a binary variable that equals 1, if a provincial party secretary or governor had substantive central experience before taking provincial positions, and 0 otherwise. In all models, manager connection consistently has significant negative effect on state ownership and state share, but the effects of leader connection are ambiguous. When the provincial party secretary has close central ties, firms are less likely to be controlled by state owners but tend to have more concentrated ownership structure if they are state owned. When the provincial governor has close central ties, firms are more likely to be controlled by state owners. The coefficients of leader connection and firm connection have different signs in three out of four models, suggesting that political connections have different impacts on firm managers and local leaders with respect to their preferences on privatization. We then include an aggregate variable firm connection, defined as the average degree of central connection of managers in total listed firms in a province, in the provincial-level test. Since the information for firm connection is only available from 2003 to 2009, we have a smaller number of observations. Firm connection has a significantly negative effect on the dynamics of privatization, indicating that firm connection does have an opposite effect on privatization progress. The coefficients on leader connection, while still having the positive sign, are no longer statistically significant. In general, these two tests suggest that political connections have different impacts on the preferences of SOEs and local leaders on privatization, but they do not necessarily reinforce or undermine each other. The results are presented in Appendix 7 and 8 available as Supplementary Material. 5. Conclusion Conventional accounts of the politics of economic development portray the major obstacle to economic reform as powerful vested interests who manipulate politicians to advance their own empires at the expense of the social interest. Whether the economic reforms can succeed or stall will depend on how the government holds off vested interests and generates support for deeper transformation. Our article provides a more nuanced perspective on vested interests for understanding the political dynamics of economic reform. Although economic reforms are often treated as a coherent package that induce political elites to form ‘special interests’ to support or oppose it, attitudes regarding the implementation of reforms are not necessarily consistent across policy areas and/or over time. Rather, different political objectives surrounding each reform should give rise to distinct strategic choices for vested interests. They may support reform in specific areas but block reforms in others, as long as they expect the reform to increase their payoffs. Although both SOE managers and local government leaders are political entrepreneurs under the nomenklura system, their different career prospects create distinct preferences for privatization. Politically connected firms are more likely to pursue rather than block privatization plans. In contrast, local government leaders, constrained by multilateral performance criteria combining economic and political objectives, have more ambiguous attitudes toward privatization. We argue that political connections play an important role in shaping the preferences of SOE managers and provincial leaders regarding privatization in China, albeit in divergent directions. On the one hand, political connections reassure SOE managers that their privileges will be preserved, therefore encouraging them to embrace privatization to maximize their profitability. On the other hand, political connections give provincial leaders more leeway to pursue noneconomic policy goals, therefore reducing their motives to use privatization to generate revenue and boost economic growth. Our multilevel empirical findings provide support to these hypotheses. At the firm level, SOEs are more likely to implement privatization plans when their managers are politically connected. At the provincial level, a province will pursue less aggressive privatization schemes when the provincial leaders have close political ties with the central government. Our results do not indicate that the Chinese experience of partial privatization is a unique model of privatization. Rather, the formation of mixed ownership consortia has increasingly become the global norm because it enables the government to implement privatization programs more easily while preserving some degree of influence in the partially privatized firms (Musaccio and Lazzarini, 2014). This trend may not only raise the conventional question of how the agency conflict between the controlling and the minority shareholders can be reduced as corporate ownership and control can be separated to the benefit of the large shareholders (Porta et al., 1999; Claessens et al., 2000), but also highlight the conflict of interest inherent in the state’s dual role as shareholder and corporate governance regulator, which can have unintended consequences well beyond potential corporate mismanagement (Pargendler, 2012). Partial privatization might be an effective arrangement to reduce resistance to reforms, but, with the potential mismatch of ownership and control, it may be more difficult to establish hard budget constraints for firms. Despite the sweeping and widespread SOE restructuring, SOEs remain the key vehicle for policy-driven investments, which are the major contributor to China’s rapidly rising corporate debt since 2008 (IMF, 2016). Understanding the political logic of the peculiar nature of partial privatization also contributes to the debate about China’s future, which may depend on the ability of the Chinese government to implement the needed reforms (e.g. Shambaugh, 2013, 2016). In recent years, the Chinese government has increasingly relied on the ‘top-level design’ (dingceng sheji) to launch important reforms, aiming to overcome strong resistance of vested interests. But various self-seeking state agencies form contingent interests through the myriads of political and business calculations, leading to inconsistent implementation of reform agendas. With the increasing embeddedness and declining autonomy for policymakers, the once well-performing developmental state models now face serious challenges. In the absence of collective responsibility, politically powerful groups try to manipulate economic reforms for their own purposes rather than for structural transformation. As such, China is falling short of its own objectives for reform. Even a Chinese state think tank has admitted that China’s state capacity for policy implementation has been weakened, indicated by the reform stalemate since 2012 (Buckley, 2017). From a broader perspective, it also helps explain why some countries or economies may be stuck in traps with little or no reform, but also indicates ways to break out of them. It suggests that when designing major economic reforms, particularly concerning privatization, reassuring vested interests is necessary for building a new base of political support for economic reforms. Over the long run, implementing economic reforms requires not only breaking up the old equilibrium by weakening supporters of the status quo, but also consolidating a new equilibrium that institutionalizes the new bases of support that have emerged. Supplementary material Supplementary material is available at Socio-Economic Review Journal online. Acknowledgements The research for this article is partially supported by the Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning. I am grateful to Zhang Jianjun and Victor Shih for generously sharing their data. I thank Jorgen Delman, Li Hui, Tang Min, Tang Shiping, Zhu Jiangnan and participants of seminars at Fudan University, University of Hong Kong, University of Copenhagen and “Five-corners-field School” for their helpful comments. I also thank Pippa Morgan for excellent research assistance. Footnotes 1 According to the World Bank, the proceeds generated from 37 large privatization transactions in China accounted for almost one-third of global privatization revenues in 2000–2008 (Nellis, 2012). 2 State shares refer to equities held by governmental agencies or authorized institutions on behalf of the state. Legal-person shares refer to equities owned by companies or institutions with a legal-person status. 3 The general rule is that shareholders owning 5% or more of a company’s outstanding shares can sell their nontradable shares after two years, whereas owners of less than 5% can sell after one year. 4 The performance evaluation guidelines include general and industry-specific targets. General targets included total profit and rate of return on equity, which account for 30 and 40% of the performance score, respectively. The industry-specific targets account for 30% of the performance score (State Council, 2012). 5 For example, Wang Min, who spent his early years as a university professor, had a reputation of promoting rapid privatization while serving as the mayor of Suzhou. He set a record of restructuring 1034 SOEs in one and a half years, which has helped his promotion to the top provincial leader in SOE-heavy provinces of Jilin and Liaoning (Caijing Magazine, 2016). In contrast, Li Keqiang, who was parachuted from the secretary of Chinse Communist Youth League to serve as party secretary of Henan and then Liaoning provinces before moving up to the central government, has not been seen as a particularly strong supporter of privatization in his tenure as a provincial leader. 6 State-invested enterprises (SIEs) consist of four types of enterprises including state-owned enterprises (guoyou duzi qiye), state-owned companies (guoyou duzi gongsi), state-controlled shareholding companies (guoyou ziben konggu gongsi) and state-invested shareholding companies (guoyou ziben cangu gongsi). See Xinhua (2014). 7 State-owned legal-person shares refer to shares directly held by SOEs. 8 The list of SASAC-supervised central SOEs is available at http://www.sasac.gov.cn/n1180/n1226/n2425/. The list of central SOEs supervised by other government agencies was collected from various websites. 9 The list of provincial SOEs was collected from the websites of provincial SASACs. 10 Bureaucratic connection is a dummy variable that equals 1 if either the chairperson of the board, the general manager, or both previously held central government positions; 0 otherwise. Legislative connection is a dummy variable that equals 1 if either the chairperson, the general manager, or both are representatives of NPC or CPPCC; 0 otherwise. 11 This method is similar to Truex’s (2014), who attempts to distinguish the possible confounding influences of formal parliament representation and informal political connections on firms’ financial performance. He finds that becoming a NPC representative will subsequently boost a firm’s profitability, suggesting that formal parliament representation has a direct causal effect on firms’ financial performance, but he also notes that firms that gained NPC representation have better political connections. Political connections are indeed built up and strengthened through executives’ previous work experiences. 12 Since all provincial leaders are deputies of the NPC, it is no longer appropriate to use NPC membership as a measure of political connection. 13 The database of provincial leaders is available at www.st360.cn/jgzyjl/ljjl. 14 The extrabudgetary account was reported separately in the official statistics before 2010, so we calculate balance using different formulas for 2003–2009 and 2010–2012. For the former, balance = ((budgetary revenue + extrabudgetary revenue) – (budgetary expenditure + extrabudgetary expenditure)) / (budgetary revenue + extrabudgetary revenue). For the latter, balance = (budgetary revenue – budgetary expenditure)/ budgetary revenue. 15 The sale of land use rights is another important channel for local governments to raise revenue. Land-related revenue accounted for a third of total local revenue at the provincial level (Lu and Sun, 2013). At the county level, land revenue contributed nearly 80% of extra-budgetary revenue (Kung and Chen, 2013). References Bai C. , Li D. , Tao Z. , Wang Y. ( 2000 ) ‘The multitask theory of State Enterprise Reform’ , Journal of Comparative Economics , 28 , 716 – 738 . Google Scholar CrossRef Search ADS Beijing Youth Daily ( 2014, August 25 ) ‘SOEs Executives Rank Top in Executive Pays of Listed Companies, Even When They Make Huge Losses’, Beijing. Biais B. , Perotti E. ( 2002 ) ‘Machiavellian Privatization’ , American Economic Review , 92 , 240 – 258 . Google Scholar CrossRef Search ADS Bortolotti B. , Milella V. ( 2006 ) Privatization in Western Europe: Stylized Facts, Outcomes, and Open Issues. Nota di Lavoro, Fondazione Eni Enrico Mattei, No. 124. Boycko M. , Shleifer A. , Vishny R. ( 1996 ) ‘A Theory of Privatization’ , The Economic Journal , 106 , 309 – 319 . Google Scholar CrossRef Search ADS Brandt L. , Li H. , Roberts J. ( 2005 ) ‘Banks and Enterprise Privatization in China’ , Journal of Law, Economics and Organization , 21 , 524 – 546 . Google Scholar CrossRef Search ADS Bremmer I. ( 2010 ) The End of the Free Market: Who Wins the War Between States and Corporations? Portfolio. Buckley C. ( 2017, March 27 ) ‘In Rare Move, Chinese Think Tank Criticizes Tepid Pace of Reform’, New York, New York Times. Caijing Magazine ( 2016, April 11 ) ‘The Rise and Fall of Wan Min (Wang Min fuchen)’. Beijing Chan V. ( 2014 ) ‘Sinopec Divests Retail Cheaply: Bad Sign for SOE Reform’, accessed at http://blogs.barrons.com/asiastocks/2014/09/15/sinopec-divests-retail-cheaply-bad-sign-for-soe-reform/ on March 15, 2016. Claessens S. , Djankov S. , Lang L. H. P. ( 2000 ) ‘The Separation of Ownership and Control in East Asian Corporations’, Journal of Financial Economics , 58 , 81 – 112 . Google Scholar CrossRef Search ADS Clarke T.s , Pitelis C. (eds) ( 2005 ) The Political Economy of Privatization . New York , Routledge . Denisova I. , Eller M. , Frye T. , Zhuravskaya E. ( 2009 ) ‘Who Wants to Revise Privatization? The Complementarity of Market Skills and Institutions’ , American Political Science Review , 103 , 284 – 304 Google Scholar CrossRef Search ADS Dinc S. , Gupta N. ( 2011 ) ‘The Decision to Privatize: Finance and Politics’ , Journal of Finance , LXVI , 241 – 270 . Google Scholar CrossRef Search ADS Du F. , Tang G. , Young M. ( 2012 ) ‘Influence Activities and Favoritism in Subjective Performance Evaluation: Evidence from Chinese State-owned enterprises’ , The Accounting Review , 87 , 1555 – 1588 . Google Scholar CrossRef Search ADS Estrin S. , Hanousek J. , Kocenda E. , Svejnar J. ( 2009 ) ‘The Effects of Privatization and Ownership in Transition Economies’ , Journal of Economic Literature , 47 , 1 – 30 . Google Scholar CrossRef Search ADS Evans P. ( 1995 ) Embedded Autonomy: States and Industrial Transformation , Princeton , Princeton University Press . Faccio M. ( 2006 ) ‘Politically Connected Firms’ , American Economic Review , March, 369 – 386 . Faccio M. , Masulis R. , McConnell J. ( 2006 ) ‘Political Connections and Corporate Bailouts’ , Journal of Finance , LXI , 2597 – 2635 . Google Scholar CrossRef Search ADS Fernandez R. , Rodrik D. ( 1991 ) ‘Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty’ , American Economic Review , 81 , 1146 – 1155 . Gan J. , Guo Y. , Xu C. ( 2012 ) What Makes Privatization Work? Evidence from a Large-Scale Nationwide Survey of Chinese Firms, Unpublished Working Paper. Hong Kong University of Science and Technology. Guo G. ( 2007 ) ‘Retrospective Economic Accountability under Authoritarianism: Evidence from China ’, Political Research Quarterly , 60 , 378 – 390 . Google Scholar CrossRef Search ADS Haber S. , Razo A. , Maurer N. ( 2003 ) The Politics of Property Rights: Political Instability, Credible Commitments, and Economic Growth in Mexico, 1876–1929 , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Haggard S. , Kaufmann R. (eds) ( 1992 ) The Politics of Economic Adjustment , Princeton , Princeton University Press . Haggard S. , Webb S. ( 1993 ) ‘What Do We Know About Political Economy of Economic Policy Reform?’ , World Bank Research Observer , 8 , 143 – 168 . Google Scholar CrossRef Search ADS Hainmueller J. ( 2012 ) ‘Entropy Balancing for Causal Effects: A Multivariate Reweighting Method to Produce Balanced Samples in Observational Studies’ , Political Analysis 20 , 25 – 46 . Google Scholar CrossRef Search ADS Hellman J. ( 1998 ) ‘Winners Take All: The Politics of Partial Reform in Postcommunist Transition’ , World Politics , 52 , 203 – 234 . Google Scholar CrossRef Search ADS Hellman J. , Jones G. , Kaufman D. ( 2003 ) ‘Seize the State, Seize the Day: State Capture, Corruption, and Influence in Transition’ , Journal of Comparative Economics , 31 , 751 – 773 . Google Scholar CrossRef Search ADS Hellman J. , Schankerman M. ( 2000 ) ‘Intervention, Corruption and Capture: The Nexus between Enterprises and the State’ , The Economics of Transition , 8 , 545 – 576 . Google Scholar CrossRef Search ADS Holz C. ( 2011 ) ‘The Unbalanced Growth Hypothesis and the Role of the State: The Case of China’s State-owned Enterprises’ , Journal of Development Economics , 96 , 220 – 238 . Google Scholar CrossRef Search ADS Huang Y. ( 2008 ) Capitalism with Chinese Characteristics: Entrepreneurship and the State , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Hung M. , Wong T. J. , Zhang T. ( 2012 ) ‘Political Considerations in the Decision of Chinese SOEs to List in Hong Kong’ , Journal of Accounting and Economics , 53 , 43 – 59 . Google Scholar CrossRef Search ADS International Monetary Fund (IMF) 2016 . Resolving China’s Corporate Debt Problem, IMF Working Paper, WP/16/203. Jefferson G. , Su J. ( 2006 ) ‘Privatization and Restructuring in China: Evidence from Shareholding Ownership, 1995–2001’ , Journal of Comparative Economics , 34 , 146 – 166 . Google Scholar CrossRef Search ADS Jia R. , Kudamatsu M. , Seim D. ( 2015 ) ‘Political Selection in China: The Complementary Roles of Connections and Performance’ , Journal of the European Economic Association , 13 , 631 – 668 . Google Scholar CrossRef Search ADS Johnson C. , Beiman I. ( 2007 ) Balanced Scorecard for State-Owned Enterprises: Driving Performance and Corporate Governance , Manila, Asian Development Bank . Kung J. , Chen T. ( 2013 ) Do Land Revenue Windfalls Reduce the Career Incentives of County Leaders? Evidence from China. Unpublished Working Paper. Hong Kong University of Science and Technology. Landry P. ( 2008 ) Decentralized Authoritarianism in China: The Communist Party’s Control of Local Elites in the Post-Mao Era , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Lardy N. ( 2014 ) Markets over Mao: The Rise of Private Business in China , Washington, DC , The Peterson Institute for International Economics . Lin L. , Milhaupt C. J. ( 2013 ) ‘We Are the (National) Champions: Understanding the Mechanisms of State Capitalism in China’ , Stanford Law Review , 65 , 697 – 760 . Li H. , Meng L. , Wang Q. , Zhou L. ( 2008 ) ‘Political Connections, Financing and Firm Performance: Evidence from Chinese Private Firms’ , Journal of Development Economics , 87 , 283 – 299 . Google Scholar CrossRef Search ADS Liao L. , Liu B. , Wang H. ( 2014 ) ‘China's Secondary Privatization: Perspectives from the Split-Share Structure Reform’ , Journal of Financial Economics , 113 , 500 – 518 . Google Scholar CrossRef Search ADS Lieberthal K. , Oksenberg M. ( 1988 ) Policy making in China , Princeton , Princeton University Press . Liu G. , Sun P. , Woo W. T. ( 2006 ) ‘The Political Economy of Chinese-Style Privatization: Motives and Constraints’ , World Development , 34 , 2016 – 2033 . Google Scholar CrossRef Search ADS Lu Y. , Sun T. ( 2013 ) Local Governments Financing Platforms in China: A Fortune or a Misfortune? IMF Working Paper, WP/13/243, International Monetary Fund. Malesky E. , Taussig M. ( 2008 ) ‘Where is Credit Due? Legal Institutions, Connections, and the Efficiency of Bank Lending in Vietnam’ , Journal of Law, Economics, and Organization , 25 , 535 – 578 . Google Scholar CrossRef Search ADS Megginson W. ( 2005 ) The Financial Economics of Privatization , New York , Oxford University Press . Google Scholar CrossRef Search ADS Mertha A. ( 2009 ) ‘Fragmented Authoritarianism 2.0’ , China Quarterly , 200 , 995 – 1012 . Google Scholar CrossRef Search ADS Milhaupt C. , Zheng W. ( 2016 ) Why Mixed-ownership Reform Cannot Fix China’s State Sector, Paulson Policy Memorandum. Ministry of Finance. ( 2008 ) China Finance Yearbook 2008 , Beijing, China Finance Press . Ministry of Finance. ( 2013 ) China Finance Yearbook 2013 , Beijing, China Finance Press . Moe T. ( 2015 ) ‘Vested Interests and Political Institutions’ , Political Science Quarterly , 130 , 277 – 318 . Google Scholar CrossRef Search ADS Montinola G. , Qian Y. , Weingast B. ( 1995 ) ‘Federalism, Chinese style: The Political Basis for Economic Success in China’ , World Politics , 48 , 50 – 81 . Google Scholar CrossRef Search ADS Murillo M. V. ( 2002 ) ‘Political Bias in Policy Convergence: Privatization Choices in Latin America’ , World Politics , 54 , 462 – 493 . Google Scholar CrossRef Search ADS Musaccio A. , Lazzarini S. ( 2014 ) Reinventing State Capitalism: Leviathan in Business, Brazil and Beyond , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Nathan A. ( 1973 ) ‘A Factionalism Model for CCP Politics’ , The China Quarterly , 53 , 34 – 66 . National Statistical Bureau . National Data, accessed at http://data.stats.gov.cn/index on April 15, 2016. Nellis J. ( 2012 ) The International Experience with Privatization: Its Rapid Rise, Partial Fall and Uncertain Future, SPP Research Paper, 5(3), The School of Public Policy, University of Calgary. Pargendler M. ( 2012 ) ‘State Ownership and Corporate Governance’ , Fordham Law Review , 80 , 2917 – 2973 . Pei M. ( 2006 ) China’s Trapped Transition: The Limits of Developmental Autocracy , Cambridge , Harvard University Press . Google Scholar CrossRef Search ADS Pepinsky T. ( 2009 ) Economic Crisis and the Breakdown of Authoritarian Regimes: India and Malaysia in Comparative Perspective , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Perotti E. ( 1995 ) ‘Credible Privatization’ , American Economic Review , 85 , 847 – 859 . Porta R. , Lopez-de-silanes F. , Shleifer A. ( 1999 ) ‘Corporate Ownership Around the World ’, Journal of Finance , LIV , 471 – 517 . Google Scholar CrossRef Search ADS Przeworski A. ( 1991 ) Democracy and the Market: Political and Economic Reforms in Eastern Europe and Latin America , New York , Cambridge University Press . Google Scholar CrossRef Search ADS Qian Y. ( 2003 ) ‘How Reform Worked in China’. In Rodrik D. (ed.), In Search of Prosperity: Analytic Narratives on Economic Growth , Princeton , Princeton University Press . Rodrik D. ( 1996 ) ‘Understanding Economic Policy Reform ’, Journal of Economic Literature , 34 , 9 – 41 . Roland G. ( 2002 ) ‘The Political Economy of Transition’ , Journal of Economic Perspectives , 16 , 29 – 50 . Google Scholar CrossRef Search ADS Shleifer A. , Treisman D. ( 2000 ) Without a Map: Political Tactics and Economic Reforms in Russia , Cambridge , MIT Press . State-owned Asset Supervision and Administration Commission (SASAC) ( 2003 ) The Interim Guidelines of Operating Performance Evaluation of Persons in Charge of Central SOEs, accessed at http://www.law-lib.com/law/law_view.asp?id=81865 on March 15, 2016. SASAC ( 2012 ) China’s State-Owned Assets Supervision and Administration Yearbook. Schamis H. ( 1999 ) ‘Distributional Coalitions and the Politics of Economic Reform in Latin America’, World Politics , 51 , 236 – 268 . Google Scholar CrossRef Search ADS Schneider B. R. ( 2004 ) ‘Organizing Interests and Coalitions in the Politics of Market Reform in Latin America’ , World Politics , 56 , 456 – 479 . Google Scholar CrossRef Search ADS Shambaugh D. ( 2013 ) China’s Goes Global: The Partial Power , New York , Oxford University Press . Shambaugh D. ( 2016 ) China’s Future , New York , Polity . Shih V. ( 2008 ) Faction and Finance in China: Elite Conflict and Inflation , New York , Cambridge University Press . Shih V. ( 2007 ) ‘Partial Reform Equilibrium: Chinese Style: Political Incentives and Reform Stagnation in Chinese Financial Policies’ , Comparative Political Studies , 40 , 1238 – 1262 . Google Scholar CrossRef Search ADS Shih V. , Jonghyuk L. , David M. ( 2016 ) Chinese Communist Party Leadership Database , San Diego, CA , UC San Diego . Shih V. , Adolph C. , Liu M. ( 2012 ) ‘Getting Ahead in the Communist Party: Explaining the Advancement of Central Committee Members in China’ , American Political Science Review , 106 , 166 – 187 . Google Scholar CrossRef Search ADS Shleifer A. , Vishny R. ( 1994 ) ‘Politicians and Firms’ , Quarterly Journal of Economics , 109 , 995 – 1025 . Google Scholar CrossRef Search ADS State Council ( 2012 ) ‘Provisional Regulations of Performance Evaluation for Central SOEs’ Leaders’, accessed at http://politics.people.com.cn/n/2013/0201/c70731-20408271.html on March 15, 2016. Truex R. ( 2014 ) ‘The Returns to Office in a ‘Rubber Stamp’ Parliament’ , American Political Science Review , 108 , 235 – 251 . Google Scholar CrossRef Search ADS Walder A. ( 2011 ) ‘From Control to Ownership: China’s Managerial Revolution’ , Management and Organization Review , 7 , 19 – 38 . Google Scholar CrossRef Search ADS Whiting S. ( 2004 ) ‘The Cadre Evaluation System at the Grass Roots: The Paradox of Party Rule’, In Naughton B. , Yang D. (eds), Holding China Together: Diversity and National Integration in the post-Deng Era , New York , Cambridge University Press . World Bank ( 2012 ) China 2030: Building a Modern, Harmonious and Creative Society , Washington, DC , World Bank . Xinhua ( 2013 ) ‘Xinhua Insight: China’s New Reforms Must Breach Vested Interests’, accessed at http://news.xinhuanet.com/english/indepth/2013-11/04/c_132857921.htm on March 15, 2016. Xinhua ( 2014 ) ‘Enterprise State-Owned Asset Law of People’s Republic of China’, accessed at http://news.xinhuanet.com/legal/2008-10/28/content_10270416.htm on March 15, 2016. Xu C. ( 2011 ) ‘The Fundamental Institutions of China’s Reforms and Development’ , Journal of Economic Literature , 49 , 1076 – 1151 . Google Scholar CrossRef Search ADS Yosuf S. , Nabeshima K. , Perkins D. ( 2006 ) Under New Ownership: Privatizing China's State-Owned Enterprises , Palo Alto , Stanford University Press . Zheng Tan Wang . accessed at www.st360.cn/jgzyjl/ljjl, on August 1, 2015. © The Author 2017. Published by Oxford University Press and the Society for the Advancement of Socio-Economics. All rights reserved. For Permissions, please email: journals.permissions@oup.com

Journal

Socio-Economic ReviewOxford University Press

Published: Jul 26, 2017

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off