Implications of Political Patronage and Political Costs for Corporate Disclosure

Abstract

We take advantage of China’s relationship-based institutional setting to investigate whether and how firms’ disclosure decision is affected by political patronage and associated political costs considerations. Using a sample of 65 firms involved in the Shanghai Pension corruption scandal of 2006, we find that relative to benchmark firms, the connected firms are associated with lower levels of disclosure prior to the scandal. However, they significantly increased their disclosures in the year immediately following the public exposure of the scandal. A content analysis indicates that the increased disclosures are value-relevant, and are not merely used as a public relations effort to subdue public outcry in the immediate aftermath of the scandal. Cross-sectional analyses further reveal that the increase in disclosure is positively associated with the extent of firm’s guanxi dependence and type/severity of involvement in the scandal. We conclude that the increased disclosures are in response to the heightened risk and potential costs of regulatory and public scrutiny in the wake of a major event involving high political and public sensitivity. The evidence is supportive of the political costs hypothesis, and has important managerial and policy implications.

Keywords

political costs corporate disclosure guanxi corruption China

Introduction

Voluntary disclosure is a central research topic in information economics with important policy implications. Early studies of motives for voluntary disclosure (e.g., Grossman & Hart, 1980; Grossman, 1981; Milgrom, 1981) argue that management has the incentive to voluntarily disclose private information to reduce information asymmetry, mitigate adverse selection risk, and maximize firm value. However, this theoretical argument has yet to receive consistent empirical support (Bamber & Cheon, 1998; Nagar, Nanda, & Wysocki, 2003). It has been argued that firms may choose not to disclose information if the benefits of disclosure are outweighed by the associated costs, such as proprietary costs (Jovanovic, 1982; Verrecchia, 1983, 1990) and political costs (Cahan, 1992; Murphy, 1996; Watts & Zimmerman, 1978, 1986) or if the disclosures are aggregate in nature (M. Rajan & Sarath, 1996).

A common problem facing empirical corporate disclosure studies is insufficient controls for endogenous relations among different aspects of the corporate information environment (Healy & Palepu, 2001). Beyer, Cohen, Lys, and Walther (2010) indicate that most of the empirical research on voluntary disclosure does not fully incorporate the costs and benefits of voluntary disclosures. They argue that institutional setting plays an important role in affecting firms’ motivation for disclosure and the costs and benefits associated with the disclosure. China’s institutional setting, in particular, the strong influence of political factors in standard setting and corporate management (Xiao, Weetman, & Sun, 2004), provides a good opportunity to test the implications of institutional and organizational realities for corporate disclosure strategy and practice.

Although China has made a significant progress in its transition to a market-based economy, non-market forces still play an important role in market activities and government/political interference continues to exert a considerable influence on resource allocation. As in other economies (e.g., Faccio, Masulis, & McConnell, 2006; Khwaja & Mian, 2005), politically connected companies in China are often given preferential treatment in terms of access to bank loans, award of contracts, government subsidies, and even bias in the judicial process (Bailey, Huang, & Yang, 2011; X. Chen, Lee, & Li, 2008; Firth, Rui, & Wu, 2011). To access scarce resources, companies have to cultivate good guanxi (personal relationships or connections) with key government officials (J. J. Chen, Cheng, Gong, & Tan, 2014). Once political ties are established, companies take great care not to expose them to outsiders, to avoid public scrutiny, regulatory investigation, and imitation by rivals. Thus, it is difficult to determine ex ante the existence of political patronage, and how it affects the corporate disclosure decision.¹

The notion that firms make financial reporting decisions to minimize political cost has been tested in the U.S. market setting. Han and Wang (1998), Patten and Trompeter (2003), and Mitra and Crumbley (2003) find that firms manage their reported earnings to avoid or minimize their exposure to political or regulatory scrutiny following major political or environmental events. Baker (1999) finds that firms that face potential political costs related to the reasonableness of their executives’ pay tend to manage their information disclosure to minimize public scrutiny. In a similar context, Murphy (1996) and Yermack (1998) also find support for the political cost hypothesis. Lev (1996) argues that earnings manipulation can occur in equilibrium only if alternative public information sources that are contemporaneously available do not suffice to undo the manipulation. One situation where manipulation cannot be undone occurs in research settings where information is not publicly available at the time reports are being managed or manipulated, but subsequently becomes publicly available, for example, when an exogenous event occurs.

The highly publicized Shanghai Pension corruption scandal of 2006 in China provides a natural experiment for investigating the impact of political patronage and associated political cost considerations on firms’ voluntary disclosure decision. Before the corruption case was exposed, the connected companies enjoyed good guanxi with the government officials, and, as a result, would face high costs (relative to benefits) if they disclosed large amounts of information that could expose their guanxi and implicate their political patrons. However, after the corruption case was brought to light and the corrupt officials were arrested, the illicit relationships are exposed and lost, and the risk of regulatory and public scrutiny rises, especially in the immediate aftermath of the event. Unless such scrutiny is averted, significant political and related costs may result, including loss of political promotion for the executives,² or worse, actual criminal prosecution. In this study, we aim to demonstrate that, in China’s highly politicized environment, firms’ voluntary disclosure decisions are, indeed, influenced by political patronage and political cost considerations.

Using a sample of 65 Shanghai-listed companies involved in the Shanghai Pension corruption scandal (“the connected firms”), we find that the connected firms disclosed less information than comparable firms before losing their political patronage. More importantly, relative to the benchmark firms as well as their own pre-event disclosure levels, the connected firms significantly increased their amounts and content of disclosure immediately after the corruption scandal was publicized and the corrupt officials arrested. This conclusion holds whether the benchmarks are firms matched on key financial variables and industry affiliation (irrespective of listing location), or are non-connected firms headquartered in Shanghai, and is robust to different measures of voluntary disclosure and different estimation windows.

A detailed content analysis shows that the increased disclosures are not limited to “boiler-plate” disclaimers. Instead, many are value-relevant information items, such as disclosures about current and planned future investments. Further cross-sectional analyses reveal that the increase in disclosure is positively associated with the extent of firm’s dependence on guanxi and firm’s type/severity of involvement in the scandal. Such cross-sectional variations are consistent with the notion that the connected firms increased their voluntary disclosures to compensate for the loss of guanxi in the scandal, and to reduce the heightened risk and associated costs of public and regulatory scrutiny in the immediate aftermath of the high-profile scandal. Our analysis of press releases and news articles for a random sample of the connected firms also suggests that the increased disclosures in the Management Discussion and Analysis (MD&A) section of the financial reports are not merely used as a public relations, “damage control” effort to subdue public outcry in the immediate aftermath of the scandal.

This article contributes to the extant literature in several ways. First, we provide convincing evidence from a natural experiment consistent with the political costs hypothesis in the context of corporate disclosure in China. The evidence indicates that when the costs of increased disclosure are high relative to the benefits, firms enjoying political patronage refrain from disclosing more information. However, when the risk of regulatory scrutiny increases in the wake of a major event involving high political and public sensitivity, these firms increase their disclosures to lower the political costs associated with (perceived) opaqueness. Our findings extend prior U.S. studies of the political costs hypothesis (Baker, 1999; Cahan, 1992; Han & Wang, 1998; Murphy, 1996) to a new and important context, namely, voluntary narrative disclosures in the world’s largest emerging market (China). Despite having very different institutional and market environments, China’s listed firms appear to adopt a disclosure strategy that is similar to that used by U.S. firms.

Second, in contrast to other empirical corporate disclosure studies that do not adequately control for endogenous relations among different aspects of the corporate information environment (Healy & Palepu, 2001), our use of a largely exogenous event, as well as a difference-in-differences research design, gives us more confidence in drawing relatively unambiguous conclusions from our results. The evidence that political ties and associated political costs considerations affect the corporate disclosure strategy and practice offers a new explanation for some well-documented findings in prior studies. Jin and Myers (2006) provide worldwide evidence that lack of transparency increases stock price synchronicity and stock crash risk. In particular, they find that China has the highest level of stock price synchronicity. Piotroski, Wong, and Zhang (2015) find that Chinese firms have significantly greater stock crash risk than the global average documented in Jin and Myers (2006), consistent with Chinese firms’ incentives to suppress the release of bad news.³ Our evidence suggests that the aforesaid findings in prior studies could be explained by the prevalence of political connections among China’s listed firms and their incentives to limit corporate disclosures.

Third, we contribute to the literature and policy debate on the effects of political ties on corporate behavior and related efficiency implications. Our results (and other studies) indicate that Chinese firms derive one of their competitive advantages from building strong guanxi with government officials. As the politically connected firms are not necessarily more efficient or profitable (our preliminary evidence is consistent with this), any preferential treatment they derive from political connections may represent resource misallocation. Severing the ties between firms and their political patrons forces the firms to become more transparent, and potentially increases stock price efficiency and allocative efficiency (Durnev, Morck, & Yeung, 2004; Wurgler, 2000). It behooves China’s policy-makers to reduce the government’s and politicians’ interference in economic activities, and allow the market to play a more effective resource-allocation role. Given that political connections and government interference are also prevalent in many transitional and even developed economies (Chaney, Faccio, & Parsley, 2011; Cooper, Gulen, & Ovtchinnikov, 2010; Faccio, 2006; Guedhami, Pittman, & Saffar, 2014), our China evidence is believed to have a wider general relevance.

The “Literature Review, Institutional Background, and Hypothesis” section reviews the relevant literature, discusses China’s institutional environment, and develops the hypothesis. The section titled “Shanghai Pension Scandal” introduces the Shanghai Pension corruption case. The section “Research Design and Methodology” discusses the empirical method and sample data. The section “Empirical Results” presents and discusses the empirical results. The section “Summary and Conclusions” concludes.

Literature Review, Institutional Background, and Hypothesis

Voluntary disclosure is the self-serving information disclosure through which a firm aims to reveal its core strategic resources to external investors and reduce the probability of adverse selection, thereby helping investors to evaluate the firm’s value more accurately. Absent any other considerations, managers are expected to fully reveal their private information (Grossman, 1981; Milgrom, 1981). However, in practice full disclosure of private information is not observed. One possible reason for incomplete disclosure is that firms, in real life, may incur disclosure costs.⁴ Thus, managers’ decisions on whether to disclose value-relevant information (and how much to disclose) are influenced by the conflicting objectives of reducing information asymmetry through disclosure and avoiding the costs of disclosing sensitive information (Healy & Palepu, 2001).

In this article, we focus on testing the political costs hypothesis in the context of corporate disclosure. The political costs hypothesis states that companies at risk of increased political scrutiny or regulatory investigations will make accounting choices that reduce such risk and the associated costs, such as wealth transfer in the form of regulation or other political/government actions (Watts & Zimmerman, 1978, 1986). Prior studies (e.g., Cahan, 1992; Cahan, Chavis, & Elmendorf, 1997; Han & Wang, 1998; Patten & Trompeter, 2003; Mitra & Crumbley, 2003) find that firms manage their reported earnings to avoid or minimize their exposure to political or regulatory scrutiny during times of high public sensitivity, such as following major political or environmental events. In the context of executive stock options reporting, Lewellen, Park, and Ro (1995), Murphy (1996), Yermack (1998), and Baker (1999) find that firms make reporting choices to minimize public scrutiny and criticism over pay levels.

Despite the fact that China has become one of the largest economies in the world and is unique on many dimensions and may not easily fit into the existing Western economic theories (Ke, Rui, & Yu, 2012), there is a dearth of research on factors influencing corporate voluntary disclosures in China. A distinctive feature of business reality in China is a high level of dependence on guanxi, especially guanxi with key resource allocators such as government officials (J. J. Chen et al., 2014). It is believed that guanxi can be a substitute for formal institutional support and is used to access scarce resources (Peng & Heath, 1996; Xin & Pearce, 1996) and enforce contracts (Chow, 1997). It also enables companies to overcome institutional barriers and instability in the face of regulatory changes, an issue that is particularly pertinent in transition economies such as China (Luo, 2003; Park & Luo, 2001). Therefore, a good guanxi network becomes a company’s competitive advantage (Chow, 1997; Park & Luo, 2001). From the perspective of the corporate executives as well, it is advantageous for their career advancement to have a strong political backing, as appointments to key positions in both government and enterprises in large part hinge on guanxi with, and favorable evaluation by, political leaders (Fan, Wong, & Zhang, 2007; Hung, Wong, & Zhang, 2012; H. Li & Zhou, 2005; Zhao & Zhou, 2004).

However, guanxi also has a dark side and represents potential pitfalls for firms (Gu, Hung, & Tse, 2008). Specifically, guanxi is characterized by its uniqueness and secrecy. Moreover, guanxi is very often associated with illegitimate deals, and is sometimes linked to nepotism, cronyism, and even corruption (Cai, Fang, & Xu, 2011; Day, 2003; Fisman & Wang, 2015). As such, firms backed by political patrons are reluctant to disclose large amounts of (useful) information for fear of exposing their guanxi and inviting public scrutiny or regulatory investigation.⁵ Using China data during 2007-2010, J. J. Chen et al. (2014) find a negative relation between firm value and voluntary disclosure for firms that rely heavily on guanxi; by contrast, for firms that rely less heavily on guanxi and more on other core competencies, they find a positive relation between firm value and voluntary disclosure. They argue that the moderating role of guanxi in the relation between firm value and voluntary disclosure results from firms conscientiously balancing the costs and benefits of voluntary disclosure relative to those of guanxi. More generally, R. Rajan and Zingales (1998) and Gul (2006) point out that in Asia’s relationship-based economic system, firms with political connections may suppress firm-specific information to hide expropriation activities by politicians and their cronies.

The above review of the literature leads to two predictions: (a) firms that enjoy political patronage are less likely to provide informative disclosures, given the high costs (relative to benefits) of disclosure; (b) loss of political patronage in a highly publicized scandal forces them to increase their disclosures to avoid the increased risk and costs of public scrutiny or political/regulatory sanctions. As an empirical matter, however, it is also possible that firms engulfed in a high-profile scandal may “clam up” and decrease their disclosures, in fear of self-incrimination.

We state our hypothesis (Political Costs Hypothesis) as follows:

Hypothesis 1: Ceteris paribus, the level of voluntary disclosure by companies backed by political patrons is relatively low prior to the loss of political patronage, and increases after the loss of political patronage.

The Shanghai Pension Scandal

We test the above hypothesis using a natural experiment, the Shanghai Pension Scandal, which is perhaps the biggest corruption case in modern Chinese history.⁶ The case was brought to light in early August 2006 when the central authorities sacked Zhu Junyi, Director of the Shanghai Municipal Labour and Social Security Bureau, who supervised the city’s pension funds, for lending RMB3.2 billion from pension funds to a private toll road company, Fuxi Investment Holding Co. The company used the funds to help bid for the operation of a Shanghai–Hangzhou expressway. Fuxi’s Chairman, Zhang Rongkun, was later detained.⁷ Subsequent investigations led to the arrest and conviction of dozens of government officials and business people, including Shanghai Communist Party of China (CPC) boss Chen Liangyu. Table 1 lists the convicted officials.

Table 1.

List of Government Officials Implicated in the Shanghai Pension Corruption Scandal.

Name of official	Time of regulatory investigation	Position	Crime	Sentence
Chen Liangyu	2006/2009	CPC Politburo member, and CPC Secretary of Shanghai City	Accepting bribery, abuse of power	18 years imprisonment
Qin Yu	2006/2008	Deputy Director of General Office of Shanghai City Government and Governor of Shanghai’s Baoshan District	Accepting bribery	Life imprisonment
Chen Chaoxian	2006/2011	Governor of Shanghai’s Changing District	Accepting bribery	13 years imprisonment
Zhu Junyi	2006/2007	Director of the Shanghai Municipal Labour and Social Security Bureau	Accepting bribery, misuse of public pension fund	18 years imprisonment
Ling Baoheng	2006/2010	Director, Shanghai City State Asset Management Bureau	Accepting bribery	8 years imprisonment
Sun Luyi	2006/2009	Vice Secretary of Shanghai City Government and Director of General Office of Shanghai City Government	Accepting bribery	15 years imprisonment
Wu Hongmei	2006/2010	Deputy Director, Shanghai City State Asset Management Bureau	Accepting bribery	11 years imprisonment
Lu Qiwei	2006/2007	Section Chief, Shanghai Municipal Labour and Social Security Bureau	Accepting bribery	8 years imprisonment
Zhu Wenjin	2006/2011	Section Chief, Shanghai Bureau of Housing and Land Resources Management	Illegal authorization of land appropriation and use, accepting bribery	15 years imprisonment

Note. CPC = Communist Party of China.

Source. Media reports and courts’ verdicts.

The Shanghai Pension corruption scandal may be considered just the tip of the iceberg of China’s guanxi-based society, where corporate executives are said to regularly collude with government officials to use profitable state-owned enterprises (SOEs) as private property or engage in outright theft of public funds. Since 2013, the new Chinese leadership has demonstrated a strong will to crack down on corruption by arresting and punishing thousands of corrupt officials to placate public anger over growing social inequalities. For this reason, it is expected that both the corrupt officials and the bribing companies/individuals are extremely secretive about their illicit ties and conduct. The largely exogenous nature of the Shanghai Pension corruption case provides a unique experiment to investigate the relation between political ties and corporate disclosure.⁸ Thus, despite a relatively small sample (with 65 event firms), the quality of the data is high and this, in our view, more than offsets the disadvantages of a small sample. To our knowledge, ours is the only study that uses a highly publicized political event to investigate the implications of political patronage and associated political costs considerations for corporate disclosure strategy and practice.⁹

Research Design and Methodology

Variable Construction

Dependent variable: Voluntary disclosure (VDI)

Following Muslu, Radhakrishnan, Subramanyam, and Lim (2015); Bozzolan, Rombetta, and Beretta (2009); and F. Li (2010) among others, we use sentence as the unit of analysis both because it is the smallest integral unit of text that conveys an idea or a message (Ivers, 1991), and because it is generally considered more reliable than pages or paragraphs (Hackston & Milne, 1996). We focus on sentences in the MD&A section of the annual financial report, in line with Muslu et al. (2014) and F. Li (2010).

Similar to the situation in the U.S. market, the MD&A of China’s listed companies generally includes management’s review and discussion of past operations, and contains discussions and analysis of major developments, trends, and uncertainty concerning the company’s future performance. Typical items of information included in the MD&A include sales turnover; profits; changes in net profit margins and causes thereof; major suppliers and customers; orders; product sales and inventory; major products; service offerings or new project development; current and planned investment projects and funding support for new investment projects; major contracts obtained and related schedules of performance and expected profitability; competitive position in the product markets; business goals, strategy, and implementation; and acquisition of strategic resources such as licenses obtained. Given the comprehensiveness of the information disclosure in the MD&A, it is believed to be indicative of firm’s level of overall disclosure. Operationally, the number of sentences in the MD&A is taken to be a measure of the level of corporate voluntary disclosure.¹⁰

Experimental variable: Involvement in the corruption scandal

Following Fan, Rui, and Zhao (2008) and Du and Du (2011), we determine the identity of the firms that are involved in the corruption scandal by referring to the media reports and courts’ verdicts. Specifically, a firm is considered to be involved in the corruption case if any one of the following conditions is met: The firm itself or its senior executives are involved in the scandal, the ultimate controlling shareholder (person or entity) is involved, or one or more than one of the other top-10 shareholders are involved. Using the above criteria, a total of 65 listed companies are identified as being involved in the corruption case. Among these, 10 companies had their senior executives involved. For 30 of the connected companies, the ultimate controlling shareholder is involved (the average shareholding is 41.49%). For 25 of the companies, one or more than one of the other top-10 shareholders are involved (the average shareholding is 2.93%). The industry distribution of the connected companies is reported in Table 2.

Table 2.

Industry Distribution of the Connected Firms.

Industry	Number of firms	% of firms
Power, gas, and water production and supply	3	4.615
Construction	1	1.538
Transportation and warehousing	1	1.538
Information technology	5	7.692
Wholesale and retailing	8	12.308
Real estate	11	16.923
Social service	3	4.615
Conglomerates	12	18.462
Manufacturing	21	32.308
Among which: Food and beverage	1	1.538
Textile, apparel, and wool	4	6.154
Paper and printing	1	1.538
Petroleum, chemistry, plastics	2	3.077
Electronics	1	1.538
Metal, non-metal	1	1.538
Machinery, equipment, instrument	8	12.308
Medicine, biological product	3	4.615

Note. The stock codes of the connected firms (source: Du & Du, 2011) are 600636, 600637, 600638, 600639, 600640, 600641, 600642, 600645, 600647, 600648, 600649, 600650, 600651, 600652, 600653, 600654, 600655, 600661, 600662, 600663, 600665, 600675, 600676, 600679, 600680, 600688, 600689, 600692, 600695, 600696, 600708, 600730, 600732, 600741, 600748, 600754, 600757, 600767, 600781, 600817, 600818, 600819, 600820, 600822, 600823, 600824, 600825, 600826, 600827, 600832, 600833, 600834, 600835, 600836, 600838, 600841, 600842, 600843, 600845, 600846, 600848, 600850, 600851, 600895, 601607. One firm (stock code 601607) experienced a merger in 2009 (the pre-merger stock code was 600849). The results are robust to the exclusion of this firm-year observation.

We consider the connected companies to have received the patronage of the corrupt government officials. A dummy variable, CORRUPT, is coded 1 for these companies, and 0 otherwise. We reason that before the scandal broke, the connected companies faced high costs (relative to benefits) associated with disclosure of potentially sensitive business information, and thus provided relatively limited voluntary disclosure. However, after the corruption case was publicized and the corrupt officials were arrested, the political patronage was lost, and the political costs of remaining (or being seen to remain) opaque substantially increased. Therefore, we predict that voluntary disclosure by the connected companies would increase immediately after the corruption scandal was exposed. As the connected firms may reestablish political ties after the political/public sensitivity of the scandal subsided, we do not have a clear a priori prediction about the changes in corporate disclosure in the longer run. We use a “difference-in-differences” research design to account for the possibility that confounding events or unobserved firm characteristics are responsible for changes in corporate disclosure around the time of the scandal.

Control variables

Following prior corporate governance and disclosure studies (e.g., Cheng & Courtenay, 2006; Eng & Mak 2003; Haniffa & Cooke, 2002; Lim, Matolcsy, & Chow, 2007; Nagar et al., 2003), we control for the following corporate governance variables: ownership type and concentration, independence and size of the board, and executive compensation. Following prior studies of corporate disclosure in China (e.g., Cheung, Jiang, & Tan, 2010) and of accounting information quality (Hwang & Lee, 2012), we also control for the following variables: total assets, return on assets (ROA), revenue, whether the firm raised new equity in the immediately following year, leverage, total assets growth rate, and cash flows from operations.

Data Sources and Sample

The corrupt officials were investigated and arrested between July 2006 and November 2006. According to the requirements of China Securities Regulatory Commission (CSRC), listed companies must release their annual reports by the end of April in the following year. Thus, when the connected companies released their 2005 annual reports (by the end of April 2006), the corrupt officials had not been investigated, but they had all been investigated or arrested by the time the 2006 annual reports were released. Therefore, we regard financial year 2005 as the year before the scandal broke, and financial year 2006 as the first year after the scandal broke. Our main objective is to determine whether the connected firms’ voluntary disclosures in financial year 2006 increased relative to financial year 2005. To mitigate possible problems with using only 2 years of data, we follow Fan et al. (2008) and compare the mean disclosure level during 2003-2005 (the 3 consecutive years before public exposure of the scandal) with the mean disclosure level during 2006-2008 (the 3 years after the scandal broke).

As an alternative basis of comparison for possible changes in the disclosure level of the connected firms, we identify a matching firm for each of the connected firms. Ideally, we would want to select a firm that is identical to a connected firm, except for the fact that it was not involved in (any known) corruption. However, it is unlikely that we would be able to find an exact match. Any less stringent criteria may be subject to selection bias, and thus to mitigate this concern we use different selection criteria and check whether our results are sensitive to the choice of the selection criteria. In the first matching effort, we select a listed firm that is in the same industry as the connected firm and is closest to the connected firm in terms of mean asset size and ROA during the 3 years before the scandal broke. Our second matching method follows Fan et al. (2008). Specifically, we classify other listed companies headquartered in Shanghai as “non-connected Shanghai firms.”¹¹ After deleting the financial companies and those that did not exist continuously throughout 2003-2008, we have a total of 57 non-connected Shanghai firms. For ease of discussion, from this point on we refer to the benchmark firms chosen through the first method as “matching firms” and to those chosen through the second method as “non-connected Shanghai firms.” The firms directly involved in the corruption case are called “connected firms.”

The financial and ownership data are taken from the China Centre for Economics Research (CCER) database and the China Stock Market and Accounting Research (CSMAR) database. CCER and CSMAR are two of the most often-used databases for China accounting and finance research.

Regression Model

To test our hypothesis, we estimate the following “difference-in-differences” model for the pooled, connected, and benchmark firms:

L n ({V D I}_{i t}) = β_{0} + β_{1} {C o r r u p t}_{i t} + β_{2} {P o s t}_{i t} + β_{3} {C o r r u p t}_{i t} \times {P o s t}_{i t} + β_{m} {C o n t r o l V a r i a b l e}_{i t} \times {P o s t}_{i t} + ω_{i t} .

${V D I}_{i t}$ is the voluntary disclosure level of firm i in year t, and is equal to the number of sentences in the MD&A section of the financial statement. ${C o r r u p t}_{i t}$ is a dummy variable coded 1 for a connected firm, and 0 otherwise. ${P o s t}_{i t}$ is a dummy variable coded 1 if the observation is from the period after the corruption scandal came to light, and 0 otherwise. The interaction term ${C o r r u p t}_{i t} \times {P o s t}_{i t}$ is the focus of our empirical tests. A positive and statistically significant coefficient on this interaction term indicates that, relative to the benchmark firms (and controlling for other factors), the connected firms saw a greater increase in their corporate disclosure after losing their political patronage.

It is useful to note that after adding the interaction term, the coefficient on ${C o r r u p t}_{i t}$ indicates whether the connected firms had a different level of voluntary disclosure relative to the benchmark firms (the omitted category) in the pre-event period. The coefficient on ${P o s t}_{i t}$ indicates whether the benchmark firms had different levels of voluntary disclosure in the post-event period, relative to the pre-event period.¹²

Empirical Results

Preliminary Analysis

We first summarize the results (untabulated for brevity) of some exploratory data analyses. First, a t test of difference in the annual means of the financial/governance variables in the pre-event period reveals no statistically significant difference between the connected firms and the matching firms, with the exception of bank loans (the group means differ in 2004) and executive compensation (for which the group means differ in 2004 and 2005). A comparison of these variables between the connected firms and the non-connected Shanghai firms also reveals no statistically significant difference, with the exception of board size (for which the group means differ in 2003). The same impression emerges if the comparison is based on the average of the 3 pre-event years. This indicates that the matching is properly done and we obtain a sample of benchmark firms before the event which are similar to the connected firms in terms of the chosen financial/governance variables and which differ only in their subsequent involvement in the corruption scandal.¹³

Table 3, Panel A, reports the temporal variations in disclosure for the connected and benchmark firms. The mean VDI of the connected firms rose from 63.472 in 2005 to 84.554 in 2006. The increase is statistically significant at the 1% level. In contrast, the benchmark firms did not experience a statistically significant change in disclosure during the same period. Extending the period of examination to the 3 years centered on public exposure of the corruption case (last column in Panel A of Table 3), we again find that the connected firms significantly increased their disclosures after the corruption was publicized. While the benchmark firms also observed an increase in disclosure, it is both numerically and statistically lower compared with that of the connected firms.

Table 3.

Summary and Descriptive Statistics.

Panel A. Time Variations in Disclosure Level.
	2003	2004	2005	2006	2007	2008	2003-2005	2006-2008
VDI
Connected firms	50.891	49.033	63.472**	84.554***	79.922	78.475	54.465	80.984***
Matching firms	54.154	49.262	62.954**	62.615	68.862	74.200	55.456	68.559***
Non-connected Shanghai firms	53.452	55.821	65.896*	62.562	67.010	72.472	58.390	67.348*
T-stat (a)	−0.75	−0.13	0.12	2.93***	1.54	0.51	−0.40	2.90***
T-stat (b)	−0.42	−0.87	−0.37	2.95***	1.74*	0.75	−0.99	3.11***

Panel B. Summary Statistics for Select Control Variables.
	Pre-event period 2003-2005	Post-event period 2006-2008
SIZE
Connected firms	21.461	21.669**
Matching firms	21.388	21.650**
Non-connected Shanghai firms	21.595	21.792
T-stat (a)	0.71	0.16
T-stat (b)	−1.08	−0.83
ROA
Connected firms	0.022	0.022
Matching firms	0.022	0.038*
Non-connected Shanghai firms	0.006	0.045**
T-stat (a)	−0.03	−1.65
T-stat (b)	1.33	−1.72*
LOAN
Connected firms	0.312	0.268*
Matching firms	0.252	0.302***
Non-connected Shanghai firms	0.285	0.249
T-stat (a)	2.69***	−1.65*
T-stat (b)	1.15	0.83
GOA
Connected firms	0.101	0.134
Matching firms	0.133	0.139
Non-connected Shanghai firms	0.065	0.143*
T-stat (a)	−1.01	−0.13
T-stat (b)	1.46	−0.19

Note. T-stat (a) indicates t test of difference in means between the connected firms and the matching firms. T-stat (b) indicates t test of difference in means between the connected firms and the non-connected Shanghai firms. Definition of variables: VDI is the number of sentences in MD&A. SIZE is the natural logarithm of total assets. ROA is return on assets (net income/total assets). LOAN is the amount of bank loans as a percentage of total assets. GOA is assets growth rate. VDI = voluntary disclosure. MD&A = Management Discussion and Analysis.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test). Significance level behind each annual value (Panel A) indicates test of difference relative to the previous year.

Figure 1 graphically portrays the change of disclosure levels over time for all three groups of firms. It is apparent that the connected firms significantly increased their disclosure, both relative to their pre-event level and relative to the benchmark firms, immediately after the corruption was publicized. However, the increase in disclosure slowed down beyond the second-year post-scandal, although it remained at a higher level than that of the benchmark firms.

Figure 1.

Temporal and cross-sectional comparison of disclosure level.

Moving on to the cross-sectional comparisons, the mean VDI of the connected firms in the pre-event period 2003-2005 (either in the individual years or over the 3-year period) is not statistically different from the benchmark firms. However, in the post-event period 2006-2008, the VDI of the connected firms is significantly higher (in both statistical terms and in magnitude) than that of the benchmark firms.

The above univariate comparisons are before controlling for other factors. The differences in the disclosure level (either across firms or over time) may be attributable to other confounding events and corporate developments. Indeed, Table 3, Panel B, shows that the average asset size (SIZE) for the matching firms increased from 21.388 during 2003-2005 to 21.650 during 2006-2008 (the increase is significant at the 5% level); the asset growth rate (GOA) for the non-connected Shanghai firms also increased from 0.065 during 2003-2005 to 0.143 during 2006-2008 (the increase is significant at the 10% level). In subsequent regression analyses, we adopt a difference-in-differences approach and control for changes in financial and corporate governance variables.

It is instructive to examine the change in profitability. The mean ROA of the connected firms in 2005 was 0.011, and it was 0.028 in 2006. Although the increase is statistically significant, it is significantly lower in magnitude than the increase from 0.009 (0.012) to 0.034 (0.031) experienced by the matching firms (the non-connected Shanghai firms) during the same time period. Extending the period to the 3 years before and 3 years after the corruption case, it is noted that while the connected firms’ mean ROA remained unchanged at 0.022, the mean ROA for the matching firms (the non-connected Shanghai firms) increased from 0.022 (0.006) to 0.038 (0.045). Thus, there is some evidence that relative to the benchmark firms, the profitability of the connected firms stagnated or even declined after they lost their political patronage. An alternative explanation is that the profitability of the connected firms had benefited from the political patronage before the corruption was exposed.

It is also of interest to examine the change in the sample firms’ ability to obtain bank loans. During 2003-2005, the connected firms, on average, obtained bank loans amounting to 31.2% of their total assets. However, in the 3 years after the corruption case was exposed, the average loan-to-asset ratio of these firms fell to 26.8%. By contrast, the matching firms saw their loan-to-asset ratio increase from 25.2% during 2003-2005 to 30.2% during 2006-2008. The evidence is consistent with that of Fan et al. (2008), and suggests that the connected firms suffered a relative decline in their ability to obtain bank loans after the corruption was exposed. An alternative explanation is that the connected firms were able to obtain more loans due to their political patronage before the corruption case was uncovered.

Table 4 reports the correlation coefficients between the main variables of interest. In the majority of cases, VDI is significantly correlated with the financial and governance variables, which suggests that they should be included as controls in the regressions. The variance inflation factor (VIF; untabulated) scores are all substantially below 4, suggesting that multicollinearity is not a serious problem.

Table 4.

Correlation Coefficients.

Panel A. Pooling Connected Firms and Matching Firms (n = 780).
	Ln(VDI)	CORRUPT	POST	SIZE	REV	GOA	TOP	SOE
Ln(VDI)	1
CORRUPT	−.052	1
POST	.293**	.000	1
SIZE	.425***	.021	.104***	1
REV	.079**	.098***	.026	.057	1
GOA	.121***	−.028	.030	.244***	−.004	1
TOP	−.126***	.042	−.135***	.261***	.084**	.167***	1
SOE	−.073**	−.098***	.023	−.023	−.052	.028	−.169***	1

Panel B. Pooling Connected Firms and Non-Connected Shanghai Firms (n = 732).
	Ln(VDI)	CORRUPT	POST	SIZE	REV	GOA	TOP	SOE
Ln(VDI)	1
CORRUPT	.083**	1
POST	.270***	.000	1
SIZE	.416***	−.049	.079**	1		.
REV	.127***	.025	.017	.014	1
GOA	.110***	.018	.077**	.251***	−.003	1
TOP	−.140***	−.008	−.125***	.364***	.087**	.160***	1
SOE	−.166***	−.074**	−.027	−.154***	−.121***	−.044	−.286***	1

Note. Definition of variables: Ln(VDI) is the natural logarithm of the number of sentences in MD&A. CORRUPT is a dummy variable coded 1 if the firm is involved in the corruption case, and 0 otherwise. POST is a dummy variable coded 1 if the observation is from the post-event period, and 0 otherwise. SIZE is the natural logarithm of total assets. REV is the revenue divided by beginning period total assets. GOA is the assets growth rate. TOP is the percentage of largest shareholding. SOE is a dummy variable coded 1 if the state is the dominant shareholder, and 0 otherwise. VDI = voluntary disclosure; MD&A = Management Discussion and Analysis.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test).

Baseline Regression Results

Table 5 reports the results of regressing disclosure level, the logarithm of VDI, against the experimental variable and the control variables. In Panel A, we first report separate regression results for the connected firms and the benchmark firms for the period 2005-2006. The connected firms, on average, saw a significant increase (approximately 24%) in their disclosure levels immediately after the corruption case. In contrast, the disclosure levels of the matching firms and the non-connected Shanghai firms did not increase significantly after the corruption case. A Wald test indicates that the coefficient on POST is larger for the connected firms than for the matching firms and the non-connected Shanghai firms.

Table 5.

Baseline Regression Results.

Panel A. Separate Regressions for the Connected Firms and Benchmark Firms (2005-2006).
	Dependent variable: Ln(VDI)
	Connected firms		Matching firms		Non-connected Shanghai firms
	Coefficient	t	Coefficient	t	Coefficient	t
_cons	−0.787	−0.65	0.415	0.38	1.517*	1.67
POST	0.244***	3.18	0.003	0.03	−0.081	−0.87
SIZE	0.141***	3.19	0.185***	3.50	0.109**	2.42
REV	0.060	0.86	−0.069	−0.82	0.287***	3.66
CFO	0.048	0.08	0.809	1.31	0.028	0.05
ROA	1.120	1.27	−0.550	−0.66	−0.727	−1.01
LEV	0.580**	2.11	−0.417	−1.53	−0.066	−0.79
GOA	0.126	0.60	0.053	0.28	0.267	1.57
INDRATIO	0.402	0.83	0.680	1.53	0.198	0.38
SOE	0.172*	1.79	−0.137	−1.35	0.044	0.33
COMPEN	0.080	1.02	0.016	0.23	−0.025	−1.35
DIRECTORS	−0.001	−0.04	0.002	0.04	0.029	0.73
TOP	0.023	0.09	0.461	1.54	−0.058	−0.18
INDUSTRY	Yes		Yes		Yes
Adjusted R²	.204		.199		.216
Observations	130		130		114

Panel B. Pooling the Connected Firms and Benchmark Firms (2005-2006).
	Dependent variable: Ln(VDI)
	Pooling connected firms and matching firms		Pooling connected firms and non-connected Shanghai firms
	Coefficient	t	Coefficient	t
_cons	0.579	0.73	0.775	1.23
POST	0.011	0.12	−0.019	−0.22
SIZE	0.161***	4.65	0.119***	4.10
REV	0.036	0.68	0.173***	3.59
CFO	0.380	0.84	−0.427	−1.05
ROA	0.187	0.30	−0.131	−0.25
LEV	−0.015	−0.08	−0.015	−0.19
GOA	0.170	1.14	0.218	1.62
INDRATIO	−0.346	−1.07	0.354	1.03
SOE	−0.026	−0.36	0.107	1.40
COMPEN	0.010	0.19	0.009	0.76
DIRECTORS	0.000	0.00	0.023	0.91
TOP	0.209	1.02	−0.013	−0.07
CORRUPT	−0.049	−0.56	0.041	0.49
POST×CORRUPT	0.225*	1.81	0.284**	2.37
POST× OTHER CONTROLS	YES		YES
INDUSTRY	YES		YES
Adjusted R²	.163		.219
Observations	260		244

Note. Definition of variables: Ln(VDI) is the natural logarithm of the number of sentences in MD&A. CORRUPT is a dummy variable coded 1 if the firm is involved in the corruption case, and 0 otherwise. POST is a dummy variable coded 1 if the observation is from the post-event period, and 0 otherwise. SIZE is the natural logarithm of total assets. REV is the revenue divided by beginning period total assets. GOA is assets growth rate. TOP is the percentage of largest shareholding. SOE is a dummy variable coded 1 if the state is the dominant shareholder, and 0 otherwise. DIRECTORS is the total number of directors on the board. INDRATIO is the percentage of independent directors on the board. COMPEN is the natural logarithm of cash compensation for top-3 highest paid executives. ROA is the return on assets (net income/total assets). LEV is the total debt over total assets. CFO is the net cash flow from operations. SEO is a dummy variable coded 1 if the firm has raised new equity (rights issue) in the following year. VDI = voluntary disclosure; MD&A = Management Discussion and Analysis.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test).

Columns 1 to 2 in Panel B of Table 5 report the results for the pooled sample of connected firms and matching firms. The coefficient on the interaction term POST×CORRUPT is 0.225 (significant at the 10% level). Thus, relative to the matching firms and after controlling for other financial and governance variables, the connected firms increased their disclosure level by a greater extent, approximately 23%, in the year immediately following the corruption scandal. The coefficient on CORRUPT is statistically indistinguishable from 0; hence, there is little evidence that the average level of voluntary disclosure for the connected firms in the year 2005 (before the corruption scandal broke) is different from that of the matching firms in the same year. The coefficient on POST (a dummy variable coded 1 if the observation is from 2006, and 0 otherwise) is also statistically indistinguishable from 0; hence, there is no evidence that the average disclosure level for the matching firms in 2006 is different from that in 2005. The results for CORRUPT and POST are consistent with the impression gleaned from our univariate analyses.

Columns 3 and 4 in Panel B of Table 5 report the results for the pooled sample of connected firms and non-connected Shanghai firms. The coefficient on POST×CORRUPT (0.284, significant at the 5% level) indicates that, from the year 2005 to the year 2006, the disclosure level of the connected firms increased by approximately 28% more than the non-connected Shanghai firms, which did not observe any significant increase in their disclosure level (the coefficient on POST is statistically indistinguishable from 0). The coefficient on CORRUPT is statistically insignificant; hence, there is no evidence that the connected firms have a higher disclosure level than comparable firms in the year 2005.

In untabulated analysis, we extend the period of examination to cover the 3 years preceding and 3 years following the year, in which the corruption scandal was exposed. The results indicate that all three groups of firms increased their disclosure levels in the 3 post-event years (2006-2008), relative to the 3 pre-event years (2003-2005), after controlling for firm characteristics. Furthermore, regression results for the pooled sample of connected firms and benchmark firms (analogous to Panel B) show that, while the connected firms, on average, disclosed less information (approximately 20% less) than the benchmark firms prior to the corruption scandal, after the corruption scandal came to light, the connected firms, on average, increased their disclosure level by 19% to 24% relative to the benchmark firms. These findings are consistent with those in Panels A and B, and with our hypothesis. As the results obtained using the two matching methods are qualitatively similar, in subsequent analyses we only tabulate the results for the connected firms and the matching firms (the results for the non-connected Shanghai firms are available on request).

Across the panels, the results for the control variables are generally in line with those of prior studies. For example, before the corruption case, larger firms tend to have higher levels of disclosure.¹⁴ Firms with higher revenues and more independent boards are associated with higher levels of disclosure.

In summary, the evidence suggests that while the connected firms disclosed less information during the period when they were enjoying political patronage, they significantly (in both statistical and economic terms) increased their disclosure after losing political patronage. As the increase is documented both inter-temporally (with the connected firms acting as their own control) and relative to benchmark firms that were not involved in the scandal, it is likely to reflect the connected firms’ response to the loss of political patronage and the heightened risk and possible costs of public scrutiny and political/regulatory sanctions following a high-profile corruption scandal.¹⁵

Content Analysis

To gain some insight into the details of the disclosures in the MD&A, we design a content analysis protocol (codebook) that classifies the disclosures into several broad categories, including operations (e.g., new investments, their financing, estimated profitability and progress), financials (e.g., earnings growth rate, geographical and segment disclosure, major customers and suppliers), strategy and future plans/outlook (e.g., major difficulties encountered and proposed solutions, industry trend, risk disclosures, future plans), related party transactions/guarantees provided, and other financial reporting/accounting issues (whether there is any change in the accounting rules). These disclosures, to the extent they are credible, inform investors about the firms’ current situation and future prospects, in terms of strengths, weaknesses, opportunities, and threats.

The above disclosure items are defined in the codebook, and scores (varying from 0.5 to 2 marks per subitem, depending on the perceived importance of the information to decision-makers) are assigned to each subitem.¹⁶ Two research assistants are trained to use the codebook, which is refined until a high level of inter-coder agreement is reached (Neuendorf, 2002).¹⁷

Based on the disclosure scores (the average of the two assistants’ results) for each variable, we compute the overall disclosure scores for each year during the period 2003-2008. Table 6 summarizes the results for the broad categories and the subitems. The following observations are noteworthy. First, for the majority of the information items, the disclosure score in the post-scandal period is statistically higher than that in the pre-scandal period. This is particularly true for disclosures related to current and planned investment projects, which are highly relevant to investors (Chung, Wright, & Charoenwong, 1998; McConnell & Muscarella, 1985). The overall disclosure score increased from 0.474 in the pre-scandal period to 0.577 post-scandal, and the increase is significant at 1% level.

Table 6.

Content Analysis of Corporate Disclosures in MD&A.Disclosure item.

	(1) Pre-event (2003-2005) disclosure score	(2) Post-event (2006-2008) disclosure score	T statistic for (2) − (1)
Broad Category A: Operations
Progress of key projects in the current year	0.585	0.738	3.24***
Profitability of key projects in the current year	0.2875	0.389	2.14**
Management and control issues	0.366	0.579	4.30***
Special events (e.g., restructuring, lawsuit)	0.263	0.353	1.92*
Detailed reasons and impacts of special events	0.1443	0.174	0.82
Restructuring requested by CSRC or Shanghai Municipal Government	0.026	0.020	−0.34
Broad Category B: Financials
Revenues, profits, and earnings growth	0.949	0.954	0.23
Segment reporting	0.989	0.995	0.57
Geographical reporting	0.826	0.9695	4.79***
Percentage supplies by top-5 suppliers	0.785	0.533	−5.41***
Names of top-5 suppliers	0.057	0.026	−1.54
Percentage sales for top-5 customers	0.732	0.548	−3.82***
Names of top-5 customers	0.041	0.062	0.92
Financial performance of holding company	0.956	0.882	−2.73***
Changes in basic financial data	0.804	0.593	−4.65***
Change in assets	0.311	0.764	10.02***
Change in earnings/net income	0.356	0.753	8.60***
Change in cash flows	0.299	0.733	9.49***
Broad Category C: Corporate strategy, risks, and future outlook
Challenges in operations and the proposed solutions	0.805	0.7487179	−1.33
Industry trend analysis	0.285	0.636	7.39***
Risks analysis	0.223	0.559	7.21***
Work plans for the next year	0.737	0.949	5.98***
Investment projects in the next year	0.241	0.138	−2.60***
Revenue and cost analysis	0.108	0.328	5.43***
Broad Category D: Related party transactions and guarantees provided
New financing raised	0.907	0.969	2.54**
Related party transactions	0.953	0.913	−1.61
Guarantees provided	0.815	0.574	−5.34***
Broad Category E: Financial reporting issues
Impact of change in the accounting rules on financial situation	0.3556701	0.831	10.87***
Overall disclosure score	0.475	0.578	9.58***

Note. MD&A = Management Discussion and Analysis; CSRC = China Securities Regulatory Commission.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test).

Second, the post-scandal increase in disclosures does not appear to be driven primarily by “boiler plate” disclaimers or discussions related to the firms’ involvement or role in the corruption scandal. Specifically, disclosures related to corporate restructuring requested by regulators/government did not increase after the scandal, and hence there is little evidence that the increased disclosure is “mechanical” or superficial in the sense that it is attributable to a known fact.

In summary, the content analysis indicates that the connected firms provided a larger amount of useful information after the corruption was exposed. The evidence lends further support to our baseline finding, and is consistent with the hypothesis that the connected firms disclosed less information while they were enjoying political patronage, but increased their disclosures (both in quantity and in quality/content) after losing the patronage.

Further Analyses

Extent of damage to guanxi and increase in disclosure post-scandal

If the loss of political patronage in the corruption scandal forces the connected firms to increase their transparency through increased disclosures, one may expect that firms suffering greater damage to their guanxi (as evidenced, for example, by greater stock price drops around the time of the scandal) should increase their disclosures by a greater extent. We adopt an event study method to investigate this possibility. Specifically, we compute the cumulative abnormal returns (CAR) for the sample of connected firms as well as the two groups of benchmark firms during the period beginning from the date when the first corrupt official was investigated/arrested (July 17, 2006) to the date when the last was investigated/arrested (November 4, 2006).¹⁸ Connected firms with above median CAR (i.e., less negative CAR) are classified as Low-Damage firms, and those with below median CAR (i.e., lower and more negative CAR) are classified as High Damage firms.

Table 7, Panel A, shows that the connected firms, on average, suffered a CAR of −3.777%, compared with 0.182% for the matching firms and −1.26% for the non-connected Shanghai firms. Thus, the connected firms suffered greater losses in the value than the benchmark firms (the difference is significant at conventional levels), consistent with the idea that the connected firms stood to lose (more) from involvement in the scandal.

Table 7.

Extent of Damage to Guanxi and Change in Post-Scandal Disclosure Level.

Panel A: CAR for Connected Firms and Benchmark Firms.
	Mean CAR (%)	T-stat (mean difference)		Median difference	χ² test stat (median difference)
Matching firms	0.182	−3.959*	—	0.375	−6.943**	—
Connected firms	−3.777		−2.517	−6.568		−7.200*
Non-connected Shanghai firms	−1.260	—		0.632	—

Panel B: Regression Results: Relation Between VDI and Extent of Damage to Guanxi.
	All connected firms
	Dependent variable: Ln(VDI)
	Years: 2004-2007		Years: 2003-2008
	Coefficient	t	Coefficient	t
_cons	−1.360	−1.59	−1.520**	−2.33
POST	0.186**	2.21	0.158**	2.33
SIZE	0.128***	3.92	0.142***	5.49
REV	0.070	1.38	0.079*	1.92
CFO	0.002	0.01	−0.105	−0.38
ROA	0.915	1.29	0.201	0.42
LEV	0.550***	3.04	0.434***	3.12
GOA	−0.027	−0.22	−0.026	−0.30
INDRATIO	−0.121	−0.32	−0.012	−0.04
SOE	0.169**	2.35	0.169***	2.89
COMPEN	0.155***	2.86	0.145***	3.48
DIRECTORS	−0.014	−0.61	−0.011	−0.65
TOP	0.231	1.18	0.251	1.59
HIGH-DAMAGE	−0.061	−0.72	−0.077	−1.14
POST×HIGH-DAMAGE	0.373***	3.10	0.392***	4.08
POST× OTHER CONTROLS	Yes		Yes
INDUSTRY	Yes		Yes
Adjusted R²	.327		.353
Observations	260		390

Panel C: Regression Results: Relation Between VDI and CAR.
	All connected firms
	Dependent variable: Ln(VDI)
	Years: 2006-2007		Years: 2006-2008
	Coefficient	t	Coefficient	t
_cons	0.799	0.73	0.032	0.03
CAR	−0.807***	−3.63	−0.752***	−4.06
SIZE	0.131***	3.27	0.141***	4.29
REV	0.046	0.66	0.060	1.01
CFO	−0.486	−1.02	−0.516	−1.34
ROA	0.843	0.85	−0.049	−0.08
LEV	0.373	1.61	0.225	1.20
GOA	0.179	1.38	0.036	0.36
INDRATIO	−0.760	−1.63	−0.155	−0.47
SOE	0.068	0.71	0.040	0.50
COMPEN	0.041	0.58	0.056	0.92
DIRECTORS	0.007	0.23	0.030	1.37
TOP	0.325	1.26	0.288	1.31
INDUSTRY	Yes		Yes
Adjusted R²	.293		.265
Observations	130		195

Note. Definition of variables: HIGH-DAMAGE is a dummy variable coded 1 (0) for connected firms whose CAR is below (above) the median CAR. Ln(VDI) is the natural logarithm of the number of sentences in MD&A. CORRUPT is a dummy variable coded 1 if the firm is involved in the corruption case, and 0 otherwise. POST is a dummy variable coded 1 if the observation is from the post-event period, and 0 otherwise. SIZE is the natural logarithm of total assets. REV is the revenue divided by beginning period total assets. GOA is the assets growth rate. TOP is the percentage of largest shareholding. SOE is a dummy variable coded 1 if the state is the dominant shareholder, and 0 otherwise. DIRECTORS is the total number of directors on the board. INDRATIO is the percentage of independent directors on the board. COMPEN is the natural logarithm of cash compensation for top three highest paid executives. ROA is the return on assets (net income/total assets). LEV is the total debt over total assets. CFO is the net cash flow from operations. SEO is a dummy variable coded 1 if the firm has raised new equity (rights issue) in the following year. CAR = cumulative abnormal return; VDI = voluntary disclosure; MD&A = Management Discussion and Analysis.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test).

Panel B shows the regression results after including a High-Damage dummy and its interaction with the POST dummy. The coefficient on the interaction term, our variable of interest, is positive and statistically significant for the subsamples. This indicates that those connected firms suffering greater damage to the firm value due to the loss of guanxi in the scandal increased their post-scandal disclosures by a greater extent than the connected firms that suffered lower damage. The coefficient on POST is positive and statistically significant, indicating that the Low-Damage firms also increased their disclosures in the post-scandal, relative to the pre-scandal, period.

Panel C reports the result of regressing CAR (as a continuous variable) and other control variables on the post-scandal disclosure level. The coefficient on CAR is significantly negative. Thus, the more severe the damage to the firm value (more negative CAR) in the scandal, the greater the increase in the post-scandal disclosure.

The evidence is consistent with our earlier findings, and lends further support to the notion that the connected firms increased their disclosures to compensate for the loss of guanxi in the corruption scandal.

Change in the disclosure level by type of involvement in the scandal

As previously noted, the 65 connected firms are not homogeneous: Some had their senior executives involved in the scandal, and for others the controlling shareholder or other top-10 shareholders were involved. As the severity of negative publicity and potential political cost implications for the firm may differ according to the type of involvement in the scandal, one may expect that the connected firms’ incentives to change their disclosure strategy also vary. To investigate this possibility, we separately examine the change in the disclosure level for the three different types of involvement in the scandal.

Table 8, Panel A, reports the change in the disclosure level pre- versus post-scandal for each of the three types of situations. Although the disclosure level of all three types of connected firms increased significantly after the scandal, the firms whose ultimate controlling shareholder (average shareholding = 41.49%) was involved in the scandal saw the largest increase in disclosure. Those firms whose senior executives were involved saw the next largest increase in disclosure. For those cases involving other top-10 shareholders (average shareholding = 2.93%), the connected firms saw the least increase in disclosure (the mean increase is not statistically significant though the median increase is). The univariate comparison is consistent with the notion that when the parties/people implicated in the scandal occupy more important positions in the firm (i.e., controlling shareholder and senior executives), the firms have greater incentives to increase their disclosures.

Table 8.

Change in the Disclosure Level by Type of Involvement in the Scandal.

Panel A: Pre- Versus Post-Scandal VDI by Type of Involvement.
Party involved	Mean			Median
	Pre-scandal (2003-2005)	Post-scandal (2006-2008)	Mean diff	Pre-scandal (2003-2005)	Post-scandal (2006-2008)	Median diff
Senior executives	45.400	73.800	28.400***	43.000	73.000	30.000***
Controlling shareholder	52.422	89.556	37.133***	46.000	89.000	43.000***
Other top-10 shareholders	60.543	73.572	13.029	46.000	61.000	15.000**

Panel B: Regression Result by Type of Involvement of Connected Firms.
	Dependent variable: Ln(VDI)
	Years: 2004-2007		Years: 2003-2008
	Coefficient	t	Coefficient	t
_cons	−1.374	−1.58	−1.561**	−2.39
POST	0.288***	3.93	0.240***	4.18
SIZE	0.139***	4.09	0.155***	5.86
REV	0.055	1.05	0.069*	1.65
CFO	−0.065	−0.18	−0.158	−0.57
ROA	1.197	1.65	0.306	0.63
LEV	0.579***	3.16	0.447***	3.20
GOA	−0.011	−0.09	−0.043	−0.48
INDRATIO	−0.023	−0.06	0.089	0.34
SOE	0.167**	2.16	0.154**	2.50
COMPEN	0.129**	2.32	0.121***	2.87
DIRECTORS	−0.005	−0.21	−0.005	−0.30
TOP	0.186	0.93	0.232	1.46
SENEXE	0.004	0.04	−0.045	−0.63
CONSHARE	0.081	1.19	0.078	1.45
POST×SENEXE	0.243	1.34	0.318**	2.23
POST×CONSHARE	0.272**	2.00	0.397***	3.70
POST× OTHER CONTROLS	Yes		Yes
INDUSTRY	Yes		Yes
Adjusted R²	.299		.343
Observations	260		390

Note. Definition of variables: SENIOREXE is a dummy variable coded 1 if the firm’s senior executives are involved in the scandal, and 0 otherwise. CONSHARE is a dummy variable coded 1 if the firm’s controlling shareholder is involved in the scandal, and 0 otherwise. Ln(VDI) is the natural logarithm of the number of sentences in MD&A. CORRUPT is a dummy variable coded 1 if the firm is involved in the corruption case, and 0 otherwise. POST is a dummy variable coded 1 if the observation is from the post-event period, and 0 otherwise. SIZE is the natural logarithm of total assets. REV is the revenue divided by the beginning period total assets. GOA is the assets growth rate. TOP is the percentage of largest shareholding. SOE is a dummy variable coded 1 if the state is the dominant shareholder, and 0 otherwise. DIRECTORS is the total number of directors on the board. INDRATIO is the percentage of independent directors on the board. COMPEN is the natural logarithm of cash compensation for top-3 highest paid executives. ROA is the return on assets (net income/total assets). LEV is the total debt over total assets. CFO is the net cash flow from operations. SEO is a dummy variable coded 1 if the firm has raised new equity (rights issue) in the following year. VDI = voluntary disclosure; MD&A = Management Discussion and Analysis.

***

, **, and * significant at the .01, .05, and .10 levels, respectively (two-tailed test).

Table 8, Panel B, shows the result of regressing POST and its separate interaction with SENIOREXE (a dummy variable coded 1 if the firm’s senior executives are involved in the scandal) and CONSHARE (a dummy variable coded 1 if the firm’s controlling shareholder is involved in the scandal) plus controls on the disclosure level. The regression coefficients on the main terms SENEXE and CONSHARE are statistically insignificant, indicating that in the pre-scandal period, those connected firms whose senior executives or controlling shareholders were subsequently involved in the scandal, on average, do not have higher disclosure levels than those connected firms whose involvement in the scandal is via other, less important, top-10 shareholders. However, after the scandal was uncovered, the former two types of connected firms increased their disclosures by a greater extent—the coefficient on POST×CONSHARE is positive and statistically significant in both subsamples, whereas the coefficient on POST×SENEXE is positive and statistically significant in the longer subperiod. The evidence from the multivariate regression is, thus, consistent with the univariate evidence above, and indicates that the extent of increase in post-scandal disclosure is associated with the extent of the connected firms’ involvement in the scandal, and, consequently, the severity of negative publicity and political cost implications for the firms.

Dependence on guanxi and increase in post-scandal disclosure

Chinese companies depend, to varying degrees, on guanxi in their value creation processes, and thus dependence on guanxi may affect the corporate disclosure decision. J. J. Chen et al. (2014) find that high guanxi-dependence firms (e.g., non-high-tech firms and firms headquartered in low-marketization regions) refrain from informative voluntary disclosures for fear of revealing sensitive information that may harm their guanxi. Although we have included industry controls in all our regression analyses, to further increase comparability with prior studies we next investigate the impact of guanxi dependence on post-scandal changes in disclosure. Following J. J. Chen et al. (2014), we create a dummy variable “High-Tech,” which is coded 1 for connected firms that belong to the high-tech industries (11 firms) and otherwise 0.¹⁹ We then rerun the baseline regression analysis for the connected firms by adding the dummy High-Tech and its interaction with POST.

The results (untabulated for brevity but available on request) indicate that the non-high-tech connected firms significantly increased their disclosures after the scandal. We find no evidence that the high-tech connected firms increased their post-scandal disclosures by a greater extent than the other connected firms. The evidence is consistent with the finding of J. J. Chen et al. (2014) that high-tech firms in general depend less heavily on guanxi in their value creation, and so the loss of guanxi in the corruption scandal did not have as significant an impact on their post-scandal disclosure decisions as it did on the other connected firms that depend more heavily on guanxi. It is reassuring that we have obtained qualitatively similar results when guanxi dependence is proxied by the extent of stock price drops around the time of the scandal, and when it is proxied by industry classification.²⁰

Timeliness of the increased disclosures

We have documented significant increases in the connected firms’ disclosure level immediately after the corruption scandal came to light. One might argue that the increased disclosure in the year immediately following the scandal was simply a public relations “damage control” mechanism, in response to the significant increase in public outcry and/or scrutiny that the scandal produced. While this alternative explanation appears consistent with the empirical evidence and cannot be completely ruled out, we argue that if the increased disclosure is actually driven by “damage control” and an effort to subdue public outcry, then timely press releases (or biased news articles via their media connections) issued in the immediate aftermath of the scandal (in July-December 2006) would be more effective than waiting several months till the annual reports were released by April 2007.

We explore this possibility first by reading through all corporate announcements released by a random sample of nine connected firms during the period July 17, 2005–December 31, 2005 (the comparison period) and the period July 17, 2006–December 31, 2006 (the event period). We find no evidence that the connected firms significantly increased their number of corporate announcements during the event period relative to the comparison period. In reality, we found no specific mention in these corporate announcements of the firms’ involvement in the scandal.

We also content analyze all newspaper reports about the connected firms during the event and comparison periods.²¹ For this purpose, we utilize the text analysis method developed by Gong, Gul, and Shan (2014), who build a comprehensive dictionary of positive and negative terms, specifically for Chinese financial news.²² We find a significant increase in the number of news articles pertaining to the connected firms during the event period, relative to the comparison period. However, the net tone (defined as the fraction of positive words minus the fraction of negative words) of the news articles, on average, became less positive during the event period.²³ This is contrary to what would be expected if the connected firms tried to influence (and succeeded in influencing) the media in their own favor.

The above analyses, though not conclusive, suggest that the increased disclosure in the 2006 annual reports of the connected firms was not (merely) used as a public relations effort to subdue public outcry or avoid public scrutiny. The more plausible conclusion (especially if the evidence is read in conjunction with the evidence from the content analysis and the cross-sectional analyses) is that the increase in MD&A disclosures is attributed to the connected firms’ desire to be (seen as) more transparent following their involvement in the high-profile political scandal.

Robustness Tests

Number of words in lieu of number of sentences

Muslu et al. (2014) and Bozzolan et al. (2009) among others argue in favor of using the number of sentences, instead of the number of words as a proxy for the disclosure level. However, words as a basic unit of linguistic content (Fagan & Gençay, 2010) may still reflect the level of a firm’s disclosure in the financial statement. Therefore, for robustness we also use the number of words in the MD&A as an alternative measure of firm-level disclosure, and reestimate Equation 1 using this new measure.

In all model specifications, the coefficients on POST×CORRUPT have the expected sign, and, in the majority of cases, are positive and statistically significant at conventional levels. Thus, our conclusion is robust to this alternative measure of disclosure.

Varying the length of the pre- and post-event period

It may be argued that focusing on the 1 year immediately preceding and the 1 year immediately following the event may lead to measurement error. For instance, if change in disclosure takes place over a relatively long period of time post-event, then focusing on the 1-year horizon may not fully capture the effect of the event. However, extending the period of examination too far during the pre-event period or the post-event period is not without problems. For example, the connected companies may not have established their political ties with the corrupt officials in 2003 (3 years prior to public exposure of the scandal). Similarly, the connected firms may have reestablished new ties with other officials in 2008 (3 years after the scandal), and thus they would have reverted back to lower disclosure levels. In both cases, the effect would be to bias against rejecting the null hypothesis. This is also evident from Table 3 and Figure 1, which show that the difference in disclosure levels between the connected and the benchmark firms declines as the post-event period is lengthened. As a sensitivity check, we change the period of examination to 2 years pre- and post-event and reestimate Equation 1.

The results (untabulated) indicate that the coefficients on POST×CORRUPT are again positive and statistically significant at conventional levels. This is true whether the connected firms are pooled with the matching firms or with the non-connected Shanghai firms. Thus, the results are qualitatively similar for event windows ranging from 1 year, 2 years, and 3 years pre- and post-event.

Summary and Conclusion

Using a sample of 65 China-listed companies involved in the Shanghai Pension corruption scandal, we find that relative to benchmark firms as well as their own pre-scandal disclosure levels, the connected firms significantly increased their voluntary disclosures after losing their political patrons in the highly publicized scandal. This conclusion holds whether the benchmarks are firms matched on key financial variables and industry affiliation, or are non-connected firms headquartered in Shanghai. The results are robust to different estimation windows and different measures of corporate disclosure.

A detailed content analysis reveals that the information disclosures are value-relevant rather than boiler-plate. We find little evidence that the increased disclosures in the MD&A section of the annual reports are driven by a mere “damage control,” public relations effort to subdue public outcry, as such disclosures were not substituted, nor even complemented, by more timely press releases or more positively biased news articles in the immediate aftermath of the scandal.

Cross-sectional analyses further reveal that firms suffering greater loss in the value from involvement in the scandal increased their post-scandal disclosures by a greater extent. Specifically, firms whose controlling shareholders or senior executives are involved in the scandal, and thus are likely to face more severe adverse publicity and higher risk/costs of regulatory scrutiny unless they increase their transparency, are associated with greater increases in post-scandal disclosures. Firms that depend more heavily on guanxi in their value creation are also associated with more significant increases in their post-scandal disclosures.

Overall, the evidence indicates that the corporate disclosure decision is influenced by political patronage and political costs considerations: When the costs of disclosing more firm-specific information are high relative to the benefits, firms are less inclined to voluntarily disclose large amounts of useful information; however, when they face a potential increase in the risk of public and regulatory scrutiny following the loss of political patronage, they increase the quantity and quality of their disclosures to minimize the costs of (perceived or actual) opaqueness during a time of high political and public sensitivity.

Our results indicate that the existence of political ties (and the resultant cost–benefit trade-off faced by firms) discourages voluntary disclosure, consistent with the Political Costs Hypothesis. To the extent that politically connected firms are shielded from market discipline due to their undue advantages derived from political connectedness, and to the extent that such firms are not necessarily more efficient and may be more opaque, severing political ties and reinforcing market discipline can improve resource allocation, corporate transparency, and investor protection. In view of our case study evidence that political ties discourage higher quantity and quality corporate disclosures, and in view of arguments in other studies that opaque firms are associated with lower stock price efficiency and higher stock crash risk, future studies may investigate the direct linkages between political ties and their real effects (e.g., corporate financing and investment decisions) using larger samples in different market settings.

Footnotes

Acknowledgements

We thank an anonymous referee and the Editor-in-Chief for their constructive comments, which help improve the quality of the article.

Authors’ Note

The usual disclaimers apply.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project is supported by the National Natural Science Foundation of China (ID: 71132001), the National Social Science Foundation of China (12BGL011), and the Programme for Changjiang Scholars and Innovative Research Team in Nankai University (PCSIRT). Partial financial support is provided by the Research Grants Council of the HKSAR (PolyU 5927/09H and PolyU 5922/13H).

Notes

References

Bailey

Huang

Yang

(2011). Bank loans with Chinese characteristics: Some evidence on inside debt in a state-controlled banking system. Journal of Financial and Quantitative Analysis, 46, 1795-1830.

Baker

(1999). Options reporting and the political costs of CEO pay. Journal of Accounting, Auditing & Finance, 14, 125-145.

Bamber

Cheon

(1998). Discretionary management earnings forecast disclosures: Antecedents and outcomes associated with forecast venue and forecast specificity choices. Journal of Accounting Research, 36, 167-190.

Beyer

Cohen

D. A.

Lys

T. Z.

Walther

B. R.

(2010).The financial reporting environment: Review of the recent literature. Journal of Accounting and Economics, 50, 296-343.

Bozzolan

Rombetta

Beretta

(2009). Forward-looking disclosures, financial verifiability and analysts’ forecasts: A study of cross-listed European firms. European Accounting Review, 18, 435-473.

Cahan

S. F.

(1992). The effect of antitrust investigations on discretionary accruals: A refined test of the political cost hypothesis. Accounting Review, 61, 77-95.

Cahan

S. F.

Chavis

B. M.

Elmendorf

R. G.

(1997). Earnings management of chemical firms in response to political costs from environmental legislation. Journal of Accounting, Auditing & Finance, 12, 37-65.

Cai

Fang

C. L.

(2011). Eat, drink, firms, government: An investigation of corruption from entertainment and travel costs of Chinese firms. Journal of Law & Economics, 54, 55-78.

Chaney

P. K.

Faccio

Parsley

D. C.

(2011). The quality of accounting information in politically connected firms. Journal of Accounting & Economics, 51, 58-76.

10.

Chen

Hong

Stein

J. C.

(2001). Forecasting crashes: Trading volume, past returns, and conditional skewness in stock prices. Journal of Financial Economics, 61, 345-381.

11.

Chen

J. J.

Cheng

Gong

S. X.

Tan

(2014). Do higher value firms voluntarily disclose more information? Evidence from China. The British Accounting Review, 46, 18-32.

12.

Chen

Lee

C.-W. J.

(2008). Government assisted earnings management in China. Journal of Accounting and Public Policy, 27, 262-274.

13.

Cheng

E. C. M.

Courtenay

S. M.

(2006). Board composition, regulatory regime and voluntary disclosure. International Journal of Accounting, 41, 262-289.

14.

Cheung

Y.-L.

Jiang

Tan

(2010). A Transparency Disclosure Index measuring disclosures: Chinese listed companies. Journal of Accounting and Public Policy, 29, 259-280.

15.

Chow

G. C.

(1997). Challenges of China’s economic system for economic theory. American Economic Review, 87, 321-327.

16.

Chung

K. H.

Wright

Charoenwong

(1998). Investment opportunities and market reaction to capital expenditure decisions. Journal of Banking & Finance, 22, 41-60.

17.

Cooper

M. J.

Gulen

Ovtchinnikov

A. V.

(2010). Corporate political contributions and stock returns. Journal of Finance, 65, 687-724.

18.

Day

(2003). The Chinese art of guanxi in business. In Berenbeim

R. E.

Arvis

J.-F.

(Eds.), Fighting corruption in East Asia: Solutions from the private sector (pp. 138-146). Washington, DC: The World Bank.

19.

(2011). Political connections, involvement in corporate scandal, and audit opinions: Evidence from the Shanghai Pension Scandal. China Journal of Management, 8, 186-194. (In Chinese)

20.

Durnev

Morck

Yeung

(2004). Value-enhancing capital budgeting and firm-specific stock return variation. Journal of Finance, 59, 65-105.

21.

Dye

R. A.

(1985). Disclosure of non-proprietary information. Journal of Accounting Research, 23, 123-145.

22.

Eng

L. L.

Mak

Y. T.

(2003). Corporate governance and voluntary disclosure. Journal of Accounting and Public Policy, 22, 325-345.

23.

Faccio

(2006). Politically connected firms. American Economic Review, 96, 369-386.

24.

Faccio

Masulis

M. W.

McConnell

J. J.

(2006). Political connections and corporate bailouts. Journal of Finance, 61, 2597-2635.

25.

Fagan

Gençay

(2010). An introduction to textual econometrics. In Ullah

Giles

D. E. A.

(Eds.), Handbook of empirical economics and finance (pp. 133-153). Boca Raton, FL: Chapman and Hall/CRC.

26.

Fan

J. P. H.

Rui

O. M.

Zhao

(2008). Public governance and corporate finance: Evidence from corruption cases. Journal of Comparative Economics, 36, 343-364.

27.

Fan

J. P. H.

Wong

T. J.

Zhang

(2007). Politically connected CEOs, corporate governance, and post-IPO performance of China’s newly partially privatized firms. Journal of Financial Economics, 84, 330-357.

28.

Firth

Rui

O. M.

(2011). The effects of political connections and state ownership on corporate litigation in China. Journal of Law & Economics, 54, 573-607.

29.

Fisman

(2001). Estimating the value of political connections. American Economic Review, 91, 1095-1102.

30.

Fisman

Wang

(2015). Corruption in Chinese privatizations. Journal of Law, Economics, & Organization, 31, 1-29.

31.

Gong

S. X.

Gul

F. A.

Shan

(2014, June). Media coverage and audit pricing in China (Working paper). Hong Kong Polytechnic, Hung Hom.

32.

Grossman

S. J.

(1981). The informational role of warranties and private disclosure about product quality. Journal of Law & Economics, 24, 461-484.

33.

Grossman

S. J.

Hart

(1980). Disclosure laws and takeover bids. Journal of Finance, 35, 323-334.

34.

F. F.

Hung

Tse

D. K.

(2008). When does guanxi matter? Issues of capitalization and its dark sides. Journal of Marketing, 72, 12-28.

35.

Guedhami

Pittman

J. A.

Saffar

(2014). Auditor choice in politically connected firms. Journal of Accounting Research, 52, 107-162.

36.

Gul

F. A.

(2006). Auditors’ response to political connections and cronyism in Malaysia. Journal of Accounting Research, 44, 931-963.

37.

Hackston

Milne

M. J.

(1996). Some determinants of social and environmental disclosures in New Zealand companies. Accounting, Auditing & Accountability Journal, 9, 77-108.

38.

Han

J. C. Y.

Wang

(1998). Political costs and earnings management of oil companies during the 1990 Persian Gulf crisis. Accounting Review, 73, 103-117.

39.

Haniffa

R. M.

Cooke

T. E.

(2002). Culture, corporate governance and disclosure in Malaysian corporations. Abacus, 38, 317-349.

40.

Hayes

Lundholm

(1996). Segment reporting to the capital market in the presence of a competitor. Journal of Accounting Research, 34, 261-327.

41.

Healy

P. M.

Palepu

K. G.

(2001). Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature. Journal of Accounting & Economics, 31, 405-440.

42.

Hung

Wong

T. J.

Zhang

(2012). Political considerations in the decision of Chinese SOEs to list in Hong Kong. Journal of Accounting & Economics, 53, 435-449.

43.

Hwang

L.-S.

Lee

W.-J.

(2012). Do takeover defenses impair equity investors’ perception of “higher quality” earnings? Journal of Accounting, Auditing & Finance, 27, 325-358.

44.

Ivers

(1991). The Random House guide to good writing. New York, NY: Random House.

45.

Jin

Myers

(2006). R2 around the world: New theory and new tests. Journal of Financial Economics, 79, 257-292.

46.

Jovanovic

(1982). Truthful disclosure of information. Bell Journal of Economics, 13, 36-44.

47.

Rui

O. M.

(2012). Hong Kong stock listing and the sensitivity of managerial compensation to firm performance in state-controlled Chinese firms. Review of Accounting Studies, 17, 166-188.

48.

Khwaja

Mian

(2005). Do lenders favor politically connected firms? Rent provision in an emerging financial market. Quarterly Journal of Economics, 120, 1371-1411.

49.

Lee

Yue

(2006). Performance, growth and earnings management. Review of Accounting Studies, 11, 305-334.

50.

Lev

(1996). Discussion: Reporting choice and the 1992 Proxy Disclosure Rules. Journal of Accounting, Auditing & Finance, 11, 516-517.

51.

Lewellen

Park

(1995). Executive stock option compensation: The corporate reporting decision. Managerial and Decision Economics, 16, 633-647.

52.

(2010). The information content of forward-looking statements in corporate filings: A Naïve Bayesian Machine Learning approach. Journal of Accounting Research, 48, 1049-1102.

53.

Meng

Wang

Zhou

(2008). Political connections, financing and firm performance: Evidence from Chinese private firms. Journal of Development Economics, 87, 283-299.

54.

Zhou

(2005). Political turnover and economic performance: The incentive role of personnel control in China. Journal of Public Economics, 89, 1743-1762.

55.

(2010). The impact of product market competition on the quantity and quality of voluntary disclosures. Review of Accounting Studies, 15, 663-711.

56.

Lim

Matolcsy

Chow

(2007). The association between board composition and different types of voluntary disclosure. European Accounting Review, 16, 555-583.

57.

Lin

Gong

S. X.

(2014). How do financial newspapers present corporate information? A comparative study of 21st Century Economic News and China Securities Journal. Modern Communication, 214, 57-63. (In Chinese)

58.

Luo

(2003). Industrial dynamics and managerial networking in an emerging market: The case of China. Strategic Management Journal, 24, 1315-1327.

59.

McConnell

Muscarella

(1985). Corporate capital expenditure decisions and the market value of the firm. Journal of Financial Economics, 14, 399-422.

60.

Milgrom

(1981). Good news and bad news: Representation theorems and applications. Bell Journal of Economics, 17, 18-32.

61.

Mitra

Crumbley

D. L.

(2003). Earnings management and politically sensitive environments: Another test of corporate response to political costs. Petroleum Accounting and Financial Management Journal, 22, 1-24.

62.

Murphy

(1996). Reporting choice and the 1992 Proxy Disclosure Rules. Journal of Accounting, Auditing & Finance, 11, 497-515.

63.

Muslu

Radhakrishnan

Subramanyam

K. R.

Lim

(2015). Forward-looking MD&A disclosures and the information environment. Management Science, 61, 931-948.

64.

Nagar

Nanda

Wysocki

(2003). Discretionary disclosure and stock-based incentives. Journal of Accounting and Economics, 34, 283-309.

65.

Neuendorf

K. A.

(2002). The content analysis guidebook. Thousand Oaks, CA: Sage.

66.

Park

S. H.

Luo

(2001). Guanxi and organizational dynamics: Organizational networking in Chinese firms. Strategic Management Journal, 22, 455-477.

67.

Patten

D. M.

Trompeter

(2003). Corporate response to political costs: An examination of the relation between environmental disclosure and earnings management. Journal of Accounting and Public Policy, 22, 83-94.

68.

Peng

M. W.

Heath

P. S.

(1996). The growth of the firm in planned economies in transition: Institutions, organizations and strategic choice. Academy of Management Review, 21, 492-528.

69.

Piotroski

J. D.

Wong

T. J.

Zhang

(2015). Political incentives to suppress negative information: Evidence from Chinese listed firms. Journal of Accounting Research, 53, 405-459.

70.

Rajan

Sarath

(1996). Limits to voluntary disclosures in efficient markets. Journal of Accounting, Auditing & Finance, 11, 361-387.

71.

Rajan

Zingales

(1998). Which capitalism? Lessons from the East Asian crisis. Journal of Applied Corporate Finance, 11, 40-48.

72.

Shon

(2010). Do stock returns vary with campaign contributions? Bush vs. Gore: The Florida recount. Economics & Politics, 22, 257-281.

73.

Verrecchia

(1983). Discretionary disclosure. Journal of Accounting & Economics, 5, 179-194.

74.

Verrecchia

(1990). Endogenous proprietary costs through firm interdependence. Journal of Accounting & Economics, 12, 245-250.

75.

Verrecchia

(2001). Essays on disclosure. Journal of Accounting & Economics, 32, 97-180.

76.

Watts

Zimmerman

(1978). Toward a positive theory of the determination of accounting standards. Accounting Review, 53, 112-134.

77.

Watts

Zimmerman

(1986). Positive Accounting Theory. Englewood Cliffs, NJ: Prentice Hall.

78.

Rui

O. M.

(2012). Ownership and the value of political connections: Evidence from China. European Financial Management, 18, 695-729.

79.

Wurgler

(2000). Financial markets and the allocation of capital. Journal of Financial Economics, 58, 187-214.

80.

Xiao

J. Z.

Weetman

Sun

(2004). Political influence and coexistence of a uniform accounting system and accounting standards: Recent developments in China. Abacus, 40, 193-218.

81.

Xin

P. R.

Pearce

J. L.

(1996). Guanxi: Connections as substitutes for formal institutional support. Academy of Management Journal, 39, 1641-1658.

82.

Yermack

(1998). Companies’ modest claims about the value of CEO stock option awards. Review of Quantitative Finance and Accounting, 10, 207-226.

83.

Zhao

Zhou

(2004). Chinese organizations in transition: Changing promotion patterns in the reform era. Organization Science, 15, 186-199.