Board characteristics and Chinese bank performance

Journal of Banking & Finance 37 (2013) 2953–2968 Contents lists available at SciVerse ScienceDirect Journal of Banking & Finance journal homepage: w...

Download PDF

588KB Sizes 0 Downloads 14 Views

Report

PDF Reader
Full Text

Journal of Banking & Finance 37 (2013) 2953–2968

Contents lists available at SciVerse ScienceDirect

Journal of Banking & Finance journal homepage: www.elsevier.com/locate/jbf

Board characteristics and Chinese bank performance Qi Liang a,b, Pisun Xu c,⇑, Pornsit Jiraporn d a

Department of Finance, School of Economics, Nankai University, Tianjin 300071, China China Academy of Corporate Governance, Nankai University, Tianjin 300071, China c Reiman School of Finance, Daniels College of Business, University of Denver, Denver, CO 80208-8951, USA d Great Valley School of Graduate Professional Studies, Pennsylvania State University, Malvern, PA 19355-1488, USA b

a r t i c l e

i n f o

Article history: Received 6 October 2011 Accepted 13 April 2013 Available online 3 May 2013 JEL classiﬁcation: G21 G28 G32 G34

a b s t r a c t Using a sample of 50 largest Chinese banks during the period of 2003–2010, we explore a comprehensive set of board characteristics (size, composition and functioning of the board) and analyze their impacts on bank performance and bank asset quality in China. We ﬁnd that the number of board meetings and the proportion of independent directors have signiﬁcantly positive impacts on both bank performance and asset quality while board size has a signiﬁcantly negative impact on bank performance. We ﬁnd new evidence that the degree of bank boards’ political connection is negatively correlated with bank performance and asset quality. The ﬁndings suggest that the board of directors plays a signiﬁcant role in bank governance in China. Ó 2013 Elsevier B.V. All rights reserved.

Keywords: Bank performance Bank reform Board of directors Corporate governance Chinese banking

1. Introduction In the wake of the recent ﬁnancial crisis, much more attention has been drawn to banks’ corporate governance. The Basel Committee on Banking Supervision (BCBS) (2006) points out that ‘‘effective corporate governance practices are essential to achieving and maintaining public trust and conﬁdence in the banking system, which are critical to the proper functioning of the banking sector and economy as a whole.’’ To date, most studies focus on the developed countries, while relatively little is known about the corporate governance structure and its role in the banking sector of emerging economies. The Chinese banking sector is the largest and most complex among emerging economies. At the end of 2010, bank assets in China totaled approximately US$13.8 trillion. Banks have an overwhelmingly dominant position in the Chinese ﬁnancial system, and are an extremely important engine of economic growth. Furthermore, with the huge size, Chinese banks have substantially increased their inﬂuence in the world ﬁnancial system.

⇑ Corresponding author. Tel.: +1 303 871 4228. E-mail addresses: [email protected] (Q. Liang), [email protected] (P. Xu), [email protected] (P. Jiraporn). 0378-4266/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jbankﬁn.2013.04.018

Since 2003, Chinese authorities have signiﬁcantly deepened banking reforms and dramatically reshaped the banking sector. With the accelerated transformation of Chinese banks into modern market-oriented ﬁnancial enterprises, a sound corporate governance system becomes vital, not only to improve the functioning of the banking sector, but also to ensure the success of the ongoing banking reforms. As pointed out in OECD (2006) in its ‘‘policy brief on corporate governance of banks in Asia’’: ‘‘shortcomings in corporate governance of banks, if widespread, can destabilize the ﬁnancial system and pose systematic risks to the real economy’’. The modern corporate governance practices are not only introduced but also implemented and ﬁrmly enforced in the Chinese banking sector. Among different governance mechanisms, the Chinese authorities place the board as an essential part of the bank governance reform. The China Banking Regulatory Commission (CBRC) issued ‘‘Guidelines for Board of Directors Code of Conduct of Joint Stock Commercial Banks’’ in 2005, aiming to standardize board structure and establish the boards of Chinese banks to be as strong and functional as those in developed countries. The boards of directors of Chinese banks are modeled after the practices in the Anglo-Saxon system, where the board is responsible for the operations of the banks and acts as both the monitor and advisor of banks’ top management. Under the newly adopted governance rules and

2954

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

deepened reforms, the resulting changes in the board structure, especially the effectiveness of the bank board, become open and important questions. This paper makes the attempt to comprehensively explore the role of the board of directors in the Chinese banking sector. The study aims to provide useful information about banks’ board structure in the context of the ongoing Chinese banking reforms, and more importantly investigates the effectiveness of the board in the Chinese banking sector. To that end, we use a sample of top 50 Chinese banks, accounting for more than 70% of the total assets of the Chinese banking sector from 2003 to 2010.1 We study an extensive set of board characteristics (size, composition and functioning of the board), analyze the impacts of board characteristics on bank performance and asset quality, and report new evidence on the role of the board of directors in Chinese banks. We show that board-level governance mechanisms, such as size, number of meetings, the proportion of independent directors and directors who are politically connected, have signiﬁcant impacts on bank performance in China. This study contributes to the literature in the following aspects. First, we extend the literature on bank board governance to a major emerging economy. Most of the earlier studies focus on the developed countries and document a signiﬁcant role of corporate governance in bank operations (e.g., Denis and McConnell, 2003; Levine, 2004; Macey and O’Hara, 2003; Adams and Mehran, 2003, 2012; Adams et al., 2010). This paper re-examines the role of corporate governance in the banking sector of emerging economies using a unique sample of Chinese banks. The mechanism and effectiveness of bank governance could be quite different in China as an emerging economy, due to different stages of ﬁnancial development, although the governance reforms in China borrow concepts and ‘‘best practices’’ from the developed countries. Such a new perspective on major emerging economies can provide more robust evidence against the potential data-snooping bias in the literature. Second, this paper provides a more complete picture of the board structure and its role in the Chinese banking sector. We examine a comprehensive set of board characteristics to capture different aspects of board and their impacts on bank performance. The board of directors is the ‘‘apex body’’ of an organization’s internal governance system (Fama and Jensen, 1983; Weisbach, 1988; Hermalin and Weisbach, 2003). An effective board has become increasingly important for Chinese banks. Driven by various reform efforts, bank managers have obtained greater control over business decision-making, as government intervention has declined and the banks’ ownership structure has diversiﬁed. But the market for corporate control is still underdeveloped. Hence, the board of directors might be a key mechanism to monitor and advise bank managers. Third, this study updates the Chinese banking literature. We extend a stream of existing research that documents the impacts of governance mechanisms, mainly ownership structure, on Chinese bank behavior. More importantly, we explore the most extensive time series of data on bank governance in China. The majority of previous literature studies a sample period ending in 2004. However, several signiﬁcant reform efforts were implemented or completed after 2004, such as the remarkable transformation of the ‘‘big four’’ state-owned banks to joint-stock commercial banks through recapitalization and IPOs. Accordingly, this paper explores a sample period of 2003–2010, reﬂecting new developments in the Chinese banking sector. Furthermore, the types of Chinese banks under examination in this study span a wide range, including state-owned commercial banks (SOCBs), joint-stock commercial 1 Our sample banks account for over 80% of the total assets in the Chinese banking sector in 2003. Although this ratio has declined in 2010, they still account for over 60% of the total assets.

banks (JSCBs), city commercial banks (CCBs) and rural commercial banks (RCBs), which differs from earlier studies. Finally, this paper provides new evidence on the impact of political connection on bank performance in a transitional economy. Most of the previous literature focuses on non-bank ﬁrms and the political connection of top management (e.g. Fan et al., 2007; Wu et al. (2012); Xu et al., 2011). In this paper, we examine the relationship between the politically connectedness of bank boards and performance. The empirical results show that banks with politically connected board have poorer performance and asset quality. Section 2 provides background information on the banking reform in China. Section 3 reviews the related literature on bank corporate governance and develops the main hypotheses. Section 4 presents the data on Chinese banks and the empirical methodology. Section 5 discusses the empirical results. Section 6 concludes.

2. An overview of Chinese bank governance reform The Chinese banking system has experienced fundamental structural changes and reforms along with China’s comprehensive economic reforms since 1978. The aim is to transform the sector from being state-owned, monopolistic and policy-driven to a multi-ownership, competitive and proﬁt-oriented system. The reform process can be divided into four periods: initial institutional restructuring (1979–1984); establishing four big state-owned commercial banks (1984–1994); reforming these state banks (1994–2003); ownership reform and foreign competition from 2003 onwards. Since 2003, several banking reform efforts have dramatically accelerated the transformation of the corporate governance system in the Chinese banking sector. First, a milestone of the bank governance reforms is the establishment of the China Banking Regulatory Commission (CBRC) in April 2003. The CBRC was put into force to ﬁll the need for the specialized regulation of an increasingly complex banking system. The CBRC is responsible for formulating regulations for the banking sector, enforcing the banking laws, supervising the banking sector and encouraging good corporate governance. The CBRC takes a top-down legislative approach to enhance the corporate governance system in the banking sector based on the modern corporate governance practices in the AngloAmerican systems. Second, the ‘‘big four’’ banks, Industrial and Commercial Bank of China (ICBC), China Construction Bank (CCB), Bank of China (BOC) and Agricultural Bank of China (ABOC), have successfully transformed from wholly state-owned banks to shareholding companies through recapitalization and IPOs. They are currently referred to as ‘‘state-owned commercial banks’’ (SOCB) and have emerged to be among the biggest banks in the world. ICBC became the largest bank in the world by pre-tax proﬁt at the end of 2010. In addition, the China’s Bank of Communications (BOCOM) was ofﬁcially included as the SOCBs by the China Banking Regulatory Commission (CBRC) in 2007, given its large size. Furthermore, ﬁnancial liberalization started through gradual deregulation of government driven allocation of credit and interest rate control, and most of the urban credit cooperatives and rural credit cooperatives have been restructured and consolidated into city commercial banks and rural commercial banks. During the restructuring process, new corporate governance mechanisms are put in place such as shareholder’s meetings, the board of directors, and top management oversight operations. The joint-stock commercial banks (JSCBs) also made improvements in corporate governance to be in line with newly adopted laws and governance rules. Third, there are an increasing number of Chinese banks going public. The traditional ‘‘big four’’ state-owned banks have

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

successfully made their initial public offerings (IPOs) in the Shanghai and Hong Kong Stock Exchanges with the China Construction Bank (CCB) being the ﬁrst one in October 2005 and the Agricultural Bank of China (ABOC) the most recent in July 2010. After public offerings, Chinese banks are subject to stricter market discipline and higher disclosure standards, which signiﬁcantly enhance the banks’ corporate governance structure. Fourth, the Chinese banking sector was fully opened to foreign ﬁnancial institutions in December 2006 to comply with the WTO commitments. Higher market competition places greater pressure for Chinese banks to improve the corporate governance structure. Finally, the introduction of foreign strategic investors is another driving force in the bank governance reforms. The Chinese government loosened foreign ownership ceilings to allow for more active roles of foreign strategic investors. Foreign strategic investors are expected to not only bring new long-term capital but also to transfer modern management techniques and corporate governance mechanisms. The ceiling of foreign ownership was lifted in January 2004 from 20% to 25% to encourage higher foreign ownership participation. Foreign investments peaked in 2005, covering all types of domestic banks.

3. Literature review and hypothesis development 3.1. Related literature on the role of corporate governance in the banking sector The role of the board of directors in the banking sector is underexplored even in developed countries.2 Pi and Timme (1993) use a sample of US commercial banks from 1987 to 1990 and ﬁnd that the banks’ cost efﬁciency and ROA are negatively correlated with CEOChairman duality and unrelated to block shareholder ownership or the proportion of outside directors. Booth et al. (2002) ﬁnd that when insider ownership increases, the percentage of outside directors decreases and the CEO-Chairman duality is less likely. Andres and Vallelado (2008) use a sample of 69 boards of large commercial banks from Canada, France, the UK, Italy, Spain, and the US from 1995 to 2005. They ﬁnd that bank performance has a signiﬁcantly positive relation with board meetings and an inverted U-shaped relation with board size and the proportion of outside directors. Pathan (2009) uses a sample of 212 large US bank holding companies from 1997 to 2004 and ﬁnds that small, less restrictive boards positively affect bank risk-taking. Cornett et al. (2009) examine how certain board structures affect earnings management in US bank holding companies from 1992 to 2002. They ﬁnd that board independence constrains earnings management. Adams and Mehran (2012) use a sample of 35 publicly traded BHCs in US over a period of 1986–1999 and examine the relationship between board governance and performance. They ﬁnd that board size is positively correlated with performance. The existing literature on bank corporate governance in China mainly focuses on the impact of ownership structure on bank performance. García-Herrero et al. (2009) use a panel data of 87 Chinese banks from 1997 to 2004 and ﬁnd that less concentrated banking ownership increases bank proﬁtability. Fu and Heffernan (2009) investigate the relationship between market structure and performance in China’s banking system from 1985 to 2002 and ﬁnd that the joint-stock banks have higher X-efﬁciency and proﬁtability than the state-owned banks. Jia (2009) provides evidence that lending by SOCBs has been relatively riskier but more prudent over time based on a sample of 14 Chinese banks from 1985 to 2004. Lin 2 As pointed out in Adams and Mehran (2012), ‘‘Most of studies of board effectiveness exclude ﬁnancial ﬁrms from their samples. As a result, we know very little about the effectiveness of banking ﬁrm governance.’’

2955

and Zhang (2009) conﬁrm that SOCBs are the worst performers (except the policy banks) in terms of simple measures of proﬁtability, efﬁciency and, asset quality on a panel of Chinese banks from 1997 to 2004. Berger et al. (2009) report foreign ownership and minor size have been associated with higher efﬁciency for a sample of 38 commercial banks in China over 1994–2003. Berger et al. (2010) ﬁnd that Chinese banks with more foreign ownership suffer a smaller diversiﬁcation discount or a lesser loss of proﬁts or increase in costs from diversiﬁcation using a sample of 88 banks from 1996 to 2006. One exception is Rowe et al. (2011). Using a sample of 41 banks, Rowe et al. (2011) examines the impacts of four board variables, namely board size, percentage of shares held by the directors, percentage of executive directors and independent directors, on Chinese bank performance. They ﬁnd that the percentage of executive directors in the boards has a signiﬁcantly negative impact while the percentage of shares owned by the board has a signiﬁcantly positive impact on bank performance.

3.2. Hypothesis development Previous literature points out that the board of directors might play a more important role in the banking sector than in nonregulated industries. First, the complex and opaque nature of the banking business aggravates the information asymmetry problem and requires more difﬁcult monitoring (Levine, 2004; Macey and O’Hara, 2003). Second, regulation plays a special role in the banking sector. On the one hand, the regulator could be considered an additional monitoring force. On the other hand, regulators might create more governance problems because the regulators could directly intervene with banks’ decision-making and pursue their own interests (Laeven and Levine, 2009). The primary focus of our paper is to examine the role of the board of directors in Chinese banks. As mentioned earlier, the Chinese banking sector has undergone tremendous reforms. The reform efforts have made signiﬁcant progress in diversifying ownership, reducing government interference and giving more control to managers in decision-making. As a result, agency problems changed from traditional managerial shirking to those most common in the Anglo-Saxon system. Some issues in ‘‘new’’ agency problems are the conﬂict of interests due to the separation of management from ownership, the manipulation of stock price and the expropriation of minority stockholders. However, market competition is still weak and the market for corporate control is underdeveloped. Hence, an effective internal governance mechanism becomes critical to supervise and advise managers in Chinese banks. In 2005, the CBRC issued ‘‘Guidelines for Board of Directors Code of Conduct of Joint Stock Commercial Banks’’, with an aim to standardize operations of bank boards and strengthen the internal bank governance.3 The boards of directors in Chinese banks gradually evolve to have similar structures and responsibilities as those in the Anglo-Saxon system.4 The bank boards are the decision-making units and ultimately responsible for the operation and management of banks (Article 7). The board shall have at least three independent directors if the commercial bank has registered capital of more than RMB 1 billion (Article 54). Each director is required to attend, in person, more than two thirds of board meetings annually (Article 57). One unique characteristic of Chinese banks is that the 3 Most banks interpret these documents as mandatory regulations, even if they are called guidelines. Furthermore, these guidelines set the standard for all banks of any signiﬁcant size, not just the joint stock commercial banks. 4 Chinese banks have a two-tier board structure composed of a Board of Directors and a Supervisory Board. However, previous literature shows that the role of the supervisory board is nominal and hence, the two-tier board indeed functions as the Anglo-Saxon unitary board (e.g., Tam, 1999).

2956

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

government still holds controlling ownership in large banks. However, each governance mechanism, e.g. the board of directors or ownership structure, plays a complementary role that might be effective in certain aspects or stages of agency problem solving. Therefore, our general hypothesis is: The board of directors in the Chinese banking plays an important role in supervising bank management and improving bank performance. Speciﬁcally, we examine a comprehensive set of board characteristics (size, composition and functioning of the board) that might affect directors’ incentives and abilities to effectively advise and monitor top management. 3.2.1. Board size Jensen (1993) argues that large corporate boards are less effective due to the problems of coordination, control, and ﬂexibility in decision-making and give excessive control to CEOs. Yermack (1996) and Eisenberg et al. (1998) provide support by showing that ﬁrms with small boards had superior ﬁnancial performance. However, other researchers argue that larger boards may improve ﬁrm performances by facilitating manager supervision and bringing more human capital to advise managers. Dalton et al. (1999) and Coles et al. (2008) ﬁnd that large boards positively impact ﬁrm performance, particularly for ﬁrms requiring more advising, such as complex ﬁrms that operate in multiple segments. Raheja (2005) therefore argues that ‘‘optimal board size and composition are functions of the directors’ and the ﬁrm’s characteristics’’. 3.2.2. Board composition The independence of directors on the boards5 is in fact a central theme in governance. The corporate governance literature offers no conclusive evidence on the role of independent directors. One strand of the literature argues that the presence of independent directors on the board tends to lessen the conﬂict of interests and be more effective in reducing the agency problem. Rosenstein and Wyatt (1990) show that stock prices react positively to the nomination of independent directors to the board. Klein (2002) argues that earnings quality increases with the proportion of independent directors. Nguyen and Nielsen (2010) ﬁnd that the stock price drops following the sudden death of independent directors. Another strand of the literature indicates that, while independent directors increase the quality of monitoring, they may lack sufﬁcient knowledge of ﬁrm-speciﬁc information and lead to sub-optimal decisions (i.e. Adams and Mehran, 2003; Raheja, 2005; Harris and Raviv, 2008). Hermalin and Weisbach (2003) and Coles et al. (2008) ﬁnd no statistically signiﬁcant impact of a ﬁrm’s number and/or percentage of outside directors on ﬁrm performance. Agrawal and Knoeber (1996) shows that the presence of independent boards decreases ﬁrm value. A growing body of literature examines the role of political connection in a transitional economy and documents its signiﬁcant impact on ﬁrm value (i.e. Fan et al., 2007; Wu et al., 2010; Xu et al., 2011). Firth et al. (2009) extends the literature to the Chinese banking sector and ﬁnd that political connection affects the banks’ lending decisions to the private sector. There is little work investigating the political connectedness of the board of directors with a notable exception of Boubakri et al. (2008). Using a sample of 245 non-bank ﬁrms across 41 countries, they ﬁnd that newly privatized ﬁrms remain politically-connected through boards after ownership divestiture and perform more poorly than their non-connected counterparts. A series of reforms have been introduced with the objective of transforming the Chinese banking sector from a wholly government-controlled system to a market-driven commercial 5 The independent director is deﬁned as such director ‘‘that has no other position in commercial banks, nor has any relation with the commercial bank or major shareholders it serves that may adversely affect his/her making independent and disinterested judgment.’’

system. However, politically-connected boards might reduce the beneﬁts of the reforms. Politically-connected boards allow more government interferences in banks’ decision makings and have the incentive structure oriented toward the government objectives. Banks with politically-connected boards are more likely to pursue political and social objectives at the banks’ expenses, which would be detrimental to bank performance. Foreign directors might bring new technology and modern managerial techniques, enhance corporate governance, exert better supervision and subsequently improve ﬁrm performance. Oxelheim and Randoy (2003) ﬁnd a signiﬁcantly higher value for ﬁrms that have foreign Anglo-American board members using a sample of ﬁrms with headquarters in Norway or Sweden. Berger et al. (2009) point out that one mechanism through which minority foreign ownership might increase Chinese banks’ efﬁciency is to take positions on the board and ‘‘leverage’’ the positions to monitor and improve bank management. Therefore, the presence of foreign directors on the board might improve Chinese banks’ performance and asset quality. Older directors might lack the incentive, energy and necessary knowledge to actively monitor and advise top management. Core et al. (1999) document that that percentage of outside directors who are over age 69 on the board is associated with weaker corporate governance and in turn higher executive compensation. With the fast pace of reforms and accelerated transformation in the Chinese banking sector, we expect that the same relationship between older directors and performance holds. The proportion of female directors on the board or the gender diversity of the board has recently become a theme in governance reform worldwide. Adams and Ferreira (2009) make the ﬁrst attempt to examine the role of female directors. They ﬁnd that female directors have better attendance records than male directors and are more likely to join monitoring committees. However, the evidence on how gender diversity impacts ﬁrm performance is mixed. Gender diversity improves performance in ﬁrms with weak governance, as measured by their abilities to resist takeover but decreases shareholder value in ﬁrms with strong governance. The impact of female directors on bank performance becomes an open question. 3.2.3. Board functioning Jensen (1993) argues that the CEO-Chairman duality, or the dual appointment of chairman and CEO, gives too much power to the individual, which could cause decision-making to not be in the best interests of minority shareholders. Yermack (1996) provides support that the CEO-chairman duality reduces board independence. In the banking industry, Pi and Timme (1993) ﬁnd that cost efﬁciency and return on assets are lower when a CEO serves as the COB in the US banking industry. A conﬂicting view argues that separating the roles of chairman and CEO can create problems in decision-making if the two powerful positions do not agree on strategies. Brickley et al. (1997) ﬁnd little evidence that combining or separating titles affects corporate performance. In contrast, Goyal and Park (2002) ﬁnd that the sensitivity of CEO turnover to performance is lower when titles are combined. Similarly, Adams et al. (2005) ﬁnd evidence that CEOs also holding the chairman title appear to hold greater inﬂuence over corporate decision making. Frequent board meetings may be a signal of a proactive board. The more frequent the meetings, the increased supervision of the top management, the more relevant the advisory role, which might improve ﬁrm performance. Furthermore, the complexity of the banking business requires a more active role of the board. Alternatively, the frequency of board meetings may increase in times of ﬁnancial distress or in times of controversial decisions. Vafeas (1999) shows that frequency of board meetings is negatively

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

2957

Table 1 Deﬁnition of variables. Variables

Predicted sign

Panel A: Dependent variables ROA ROE Pre-provision proﬁt ratio NPL ratio Stock of NPLs NCO ratio Level of NCOs Panel B: Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector BusyDirector ForeignDirector OldDirector FemaleDirector

Measures Net income/total assets Net income/stockholders’ equity Operating income minus operating expenses/total assets Problem loans/total loans Natural log of total problems loans Net charge-offs/total loans Natural log of (gross charge-offs minus recoveries)

variables +/ +/ +/ +/ + +/

Panel C: Other control variables BankSize LoanRatio CapitalRatio Listed ForStgInvestor PLGState PLGForeign PLGPrivate ShareOth GeoIndex

The number of directors in the board Number of board meetings Dummy variable equal 1 if CEO is also chairman of the board and 0 otherwise Percentage of directors who are independent. Percentage of directors who are politically connected Percentage of directors who serve on 3 or more other boards Percentage of directors who are foreigners Percentage of directors who are older than 70 Percentage of directors who are female Natural log of total assets Total loans/total assets Equity/total assets Dummy variable equal 1 if a bank has been listed at the end of the year and 0 otherwise Dummy variable equal 1 if a bank has foreign strategic investor at the end of the year; 0 otherwise Percentage of shares held by the largest shareholders if the largest shareholder is government or government agency Percentage of shares held by the largest shareholders if the largest shareholder is a foreign investor Percentage of shares held by the largest shareholders if the largest shareholder is a private investor Herﬁndahl index of shareholdings of the second to tenth largest shareholders Natural log of weighted average GDP per capita of cities that a bank’s branches are located

related to performance, which may be the result of boards meeting more often to address poor performance. Busy directors are often expected to be less active monitors relative to directors who sit on fewer boards. Core et al. (1999) ﬁnd that the number of busy directors is correlated with less effective corporate governance and higher CEO pay. Fich and Shivadasani (2004) document that when a majority of outside directors serve on three or more boards, ﬁrms exhibit lower market-to-book ratios and lower operating proﬁtability. Ahn et al. (2010) ﬁnd that directors serving on multiple boards allow value-destroying acquisitions. Jirapon et al. (2009) report that directors who hold more outside directorships serve on fewer board committees, which compromises a director’s ability to effectively perform monitoring duties. The opaque and complex business of banks might require more time and attention from director in order to effectively ﬁll in the monitoring and advising roles. Hence, we expect that the busy director has a negative impact on bank performance. Table 1 provides the deﬁnitions of each board variable and the predicted relationship with bank performance.

4. Data and methodology 4.1. Data Our sample is an unbalanced panel of 52 Chinese banks during the period of 2003–2010. These banks are the top 50 banks based on the total assets according to the annual China Banking Regulatory Commission (CBRC) ranking.6 Our sample includes four stateowned commercial banks (BOC, CCB, ICBC, BOCOM7), all 12 jointstock commercial banks, 30 city commercial banks and six rural 6 Our sample includes the top 50 largest banks in terms of total assets. Two banks are tied in the ranking. That is why there are a total of 52 banks. 7 The Agricultural Bank of China (ABOC) has not started the restructuring program until late 2008 and eventually went on IPO in July, 2010. Hence, there is very limited board information reported for the sample period and is not included in the sample.

commercial banks. The CBRC’s criteria are used to determine the type of each bank. The total assets of our bank sample accounted for over 60% of the total market share of the Chinese banking industry in 2010.8 Financial information was mainly obtained from the Bankscope database. Most of the sample banks follow the Chinese Accounting Standards (CAS) while some banks, including the joint ventures and banks listed in the stock market, also prepared annual reports based on the International Accounting Standards (IAS). The CAS was recently developed following the principles of IAS. Similar to Berger et al. (2009), we compare the ﬁnancial statements of the same bank which reports under both the CAS and IAS and do not ﬁnd a material difference. The data on detailed board and ownership structure was handcollected mostly from the individual banks’ annual reports9 and additionally from other sources including the annual issues of the

8 We include four of the ‘‘big ﬁve’’ banks with a market share measured by total assets of about 42% at the end of 2010. We include all 12 joint-stock commercial banks – known as the ‘‘second-tier’’ domestic banks which have a market share of about 16%. We include 30 city commercial banks, which possess almost half of the assets of all city banks and have a market share of about 4.3%. We also include 6 rural commercial banks with a market share of about 0.85%. The city and rural commercial banks are known as the ‘‘third-tier’’ banks in the industry. Our sample of the top 50 banks covers over 60% of the banking assets in China at the end of 2010. 9 In 2002, the CBRC issued a guideline to regulate the information disclosure of commercial banks (‘‘Guideline on commercial banks’ information disclosure’’). Following the guideline, commercial banks began to issue annual reports starting in 2003. As of 2007, almost all banks with total assets above $1 billion have annual reports either published on the banks’ websites or available upon request as hard copies or reprinted text documents. Some banks also published annual reports on the periodicals such as ‘‘Financial News’’, which is the ofﬁcial publication of the People’s Bank of China and serves as the media platform for the CBRC. In the guideline, the CBRC required all the banks to disclose corporate governance practices in the annual reports. Chinese commercial banks have continuously improved the disclosure quality and disclose more detailed information on bank governance over the years. However, the detailed information on board structure and director characteristics is not available for all banks in every year during our sample period. Hence, we use the unbalanced sample in the empirical analysis.

2958

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

Almanac of China’s Finance and Banking and news releases. The director characteristic variables, such as age and gender, are collected from each director’s biography. For those banks that do not report independent directors’ information, we read the directors’ biographies and made a judgment as to whether he or she is an independent director based on the CBRC criteria, which is deﬁned as ‘‘such director that has no other position in commercial banks, nor has any relation with the commercial bank or major shareholders it serves that may adversely affect his/her making independent and disinterested judgment’’. The data on the banks’ branch information is from the individual banks’ annual reports. The data on the ownership and identity of top 10 shareholders is from individual banks’ annual reports. The data on economic indicators including the city and province’s GDP per capita was from China Statistical Yearbook and China City Statistical Yearbook published by China Statistics Press. 4.2. Empirical methodology Our main model setup is speciﬁed as the following:

Bank performance or loan qualityi;t X j ¼aþ bj board variablesi;t þ c control variablesi;t þ ei;t

ð1Þ

j

where i goes from bank 1 to bank 52 and t takes the values of the years from 2003 to 2010. The b parameters capture the potential impacts of various board characteristics on bank performance and asset quality. 4.2.1. Bank performance and loan quality measures Although banking institutions have become increasingly complex, proﬁtability and loan qualities are the underlying drivers of ﬁrm performance. Following previous literature (Andres and Vallelado, 2008; Lin and Zhang, 2009; García-Herrero et al., 2009; Berger et al. 2010), we use three bank performance measures in our analysis. Return on Asset (ROA) is calculated as net income over total assets. It is a widely used measure of bank proﬁtability in the banking literature. It shows how efﬁciently a bank uses its assets to generate revenue. The pre-provision proﬁtability ratio is operating proﬁt (operating income minus operating expenses) over total assets. The ad hoc and government-directed provisioning policy could cause the problem of non-homogeneous data for provisioning across Chinese banks. The pre-provision proﬁtability ratio is less likely to be affected by such problem. Return on Equity (ROE) is net income over equity capital and is an assessment of the ﬁnancial return of shareholders’ investment.10 We use four proxies to measure banks’ asset quality: the nonperforming loan ratio (NPL ratio), the stock of NPLs, the net charge-off ratio (NCO ratio), and the level of NCOs (Lin and Zhang, 2009; García-Herrero et al., 2009).11 These variables are essentially credit quality measures with respect to the banks’ lending practices. Excess levels of low quality loans are expected to result in higher non-performing assets and net charge-offs. For NPLs, we use problem loans as deﬁned as substandard, doubtful and loss under the 10 Similar to García-Herrero et al. (2009), we argue that ROA and pre-provision proﬁt are more appropriate than the ROE when applied to the Chinese banks because bank equity is usually low and might suffer from the artiﬁcial changes due to the recapitalization program, which requires additional caution in interpreting the ROE results. 11 Loan loss reserve is another measure commonly used in the banking literature to measure asset quality. The loan loss provision ratio is an ex ante measure of risk while the non-performing loans ratio is an ex post of risk. However, similar to GarcíaHerrero et al. (2009), we ﬁnd a very low correlation between the stock of loan loss reserve and that of NPLs at 0.37 in our sample Chinese banks. Hence, we focus on nonperforming loans and net charge-offs.

5-tier classiﬁcation system,12 which was unanimously adopted by Chinese banks in 2004. The non-performing assets ratio (NPL ratio) is calculated as the amount of non-performing loans over total loans. The stock of NPLs is the natural log of the total amount of NPLs each year. The CBRC particularly uses the two measures to assess the bank’s asset quality. The net charge-offs ratio (NCO ratio) is calculated as the amount of net charge-offs over total loans. The Level of NCOs is the natural log of the total amount of net charge-offs each year. Net charge-offs represent the amount of asset write-downs minus recoveries of previous write-downs. A charge-off or asset write-down represents the removal of an account from a bank’s books as an asset after it has been delinquent for a period of time and the lender absorbs the outstanding balance as a loss. All four measures have negative correlations with a bank’s asset quality. 4.2.2. Board characteristics measures Following the previous literature mentioned in literature review section, we measure board characteristics from three aspects: board size, board composition, and board functioning. The board characteristics variables include the number of directors serving on the board (Board Size); the number of board meetings during the year (Meetings); the dual position of the chairman of the board and the president of the bank (Duality); the proportion of independent directors in the board (IndepDirector) where the independent director is deﬁned as ‘‘such director that has no other position in commercial banks, nor has any relation with the commercial bank or major shareholders it serves that may adversely affect his/her making independent and disinterested judgment’’ (CBRC); the proportion of politically-connected directors in the board (PoliticalDirector) where the politically-connected director is the director who is currently serving or formerly served in the government or military; the proportion of busy directors on the board (BusyDirector) where the busy director is deﬁned as the director who serves on three or more boards; the proportion of directors on the board who are foreigners (ForeignDirector); the proportion of directors on the board who are older than 69 years old (OldDirector); the proportion of female directors on the board (FemaleDirector).13 4.2.3. Control variables 4.2.3.1. Control for the ownership structure. Both the type of ownership and the degree of ownership control have been shown to signiﬁcantly relate to ﬁrm value (e.g. Wang, 2005; Wei et al., 2005; Chen et al., 2009). Previous banking literature focuses on the impact of ownership by type of shareholder and mainly uses the ownership dummy variables in the empirical analysis (e.g. Berger et al., 2009; García-Herrero et al., 2009; Fu and Heffernan, 2009; Jia, 2009; Lin and Zhang, 2009). We extend the banking literature and construct several continuous variables to control for both the type of ownership and the degree of ownership concentration. 12 Under the 5-tier loan classiﬁcation standard, a loan is classiﬁed as substandard when borrowers’ abilities to service their loans are in question, as doubtful when borrowers cannot pay back the principal and interest in full and signiﬁcant loss will be incurred, as loss when the principal and the interest of loan cannot be recovered or only a small portion can be recovered after taking all possible measures and necessary legal procedures. 13 Following a few recent studies that examine the impact of directors’ ﬁnancial expertise on U.S. bank performance (e.g. Güner et al., 2008; Minton et al., 2011), we explore the role of ﬁnancial experts in the Chinese banking sector. We classify a director as a ﬁnancial expert if he or she works within a ﬁnancial institution or has a ﬁnance-related role within a non-ﬁnancial ﬁrm (e.g. CFO, accountant, treasurer, VP ﬁnance) or academic institution (e.g. professor in ﬁnance, accounting, economics or business), or is a professional investor (e.g. mutual fund). We ﬁnd that on average, bank boards in China have 45% of directors who have ﬁnancial experience or background. Boards in SOCBs and JSCBs have more directors with ﬁnancial expertise than city commercial banks and rural commercial banks. We regress each performance and asset quality variable on the proportion of director with ﬁnancial expertise on the board. The empirical results show that ﬁnancial experts on the board have no signiﬁcant impact on bank performance and asset quality in China.

2959

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

The level and identity of the controlling shareholder is measured by the percentage of shares owned by the largest shareholder if the largest shareholder is the state and state-owned enterprise (PLGState), foreign investor (PLGForeign), or private investor (PLGPrivate).14 We also include a variable to measure the concentration of shares held by the second to the tenth largest shareholders (ShareOth), which captures the likelihood that other large shareholders will challenge the largest shareholder. The ShareOth is the Herﬁndahl inP Sn 2 dex deﬁned as 10 , where Sn is the number of shared held by n¼2 S the nth largest shareholder and S is the number of total outstanding shares. All ownership variables are measured at the bank-year level. 4.2.3.2. Control for the geo-economy effect. Using a ﬁeld survey on 20 banks in China, Ferri (2009) shows that the performance of city commercial banks in economically developed areas are better than those in less developed ones. To control for the potential geoeconomic location effects, we construct a geo-economic index (GeoIndex), which is the weighted average of the natural log of GDP per capita for all cities that a bank has branches in, using the number of branches to compute the weights.15 The geoeconomic index captures the regional economic environment that a bank is exposed to. We expect a positive relationship between the geo-economic index and bank performance. The geo-economic index is measured at the bank-year level. 4.2.3.3. Other control variables. Based on the existing literature (Lin and Zhang, 2009; García-Herrero et al., 2009; Berger et al., 2010) and Chinese banks’ unique characteristics, we include additional control variables that might affect bank performance. Bank size is measured as the natural log of total assets (BankSize). The loan ratio is the amount of total loans over total assets (LoanRatio). We use the ratio of equity over total assets to account for bank capitalization (CapitalRatio). A dummy variable for stock listing (Listed) is also included to account for the increasing privatization and more diffused ownership, which equals one if the bank is listed in a stock exchange and zero otherwise. In addition, we include a dummy variable which equals one if the bank has foreign strategic investors (ForStgInvestor) to account for their special roles in the Chinese banking system. Foreign strategic investors are expected to not only bring new long-term capital but also to transfer modern management techniques and corporate governance mechanisms. Table 2a Panel A reports the descriptive statistics for the dependent, independent, and control variables. The average ROA of our sample banks is 1%, ROE is 14%, pre-provision proﬁt is 1%, NPL ratio is 3% and NCO ratio is 1%. The average size of our sample Chinese bank boards is 13.8, which are smaller compared to those in developed countries, with 17.97 directors in Adams and Mehran (2008) using a sample of over 30 US banks from 1986 to 1999 and with 15.78 directors in Andres and Vallelado (2008) using a sample of 69 banks in OECD countries from 1996 to 2006. The average number of meetings per year is 7.01, which is lower compared to 8.48 in Adams and Mehran (2012) and 10.45 in Andres and Vallelado (2008). On average, 9% of Chinese banks have the dual position for president and chairman of the board. On average, bank boards in China have 23% of directors who are independent, 37% directors who are politically connected, 13% of directors who serve on more than three boards, 6% of directors who are foreigners, 1% of directors who are older than 70, and 11% of directors are female. In

14 The foreign (or private) investor refers to foreign (or private) individuals and institutions. i P hNj;i;t 15 Speciﬁcally, the geo-economic index is deﬁned as m j¼1 Ni;t GDP j;t . Ni,t is the

total number of branches of bank i in year t. Nj,i,t is the number of branches of bank i that are located in city j in year t. GDPj,t is the natural log of GDP per capital of city j in year t. m is the total number of cities in which bank i has branch(es) in year t.

Table 2a Panel A: Descriptive statistics. Variables

N

Mean

STD

Min

Max

Panel A1: Bank performance variables ROA 359 ROE 357 Pre-provision proﬁt ratio 353 NPL ratio 357 Stock of NPLs 337 NCO ratio 317 Level of NCOs 288

0.01 0.14 0.01 0.03 21.06 0.01 18.74

0.01 0.08 0.01 0.04 2.19 0.01 2.67

0.01 0.24 0.01 0.00 12.84 0.00 9.21

0.07 0.49 0.09 0.53 27.43 0.07 24.88

Panel A2: Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector BusyDirector ForeignDirector OldDirector FemaleDirector

13.80 7.01 0.09 0.23 0.37 0.13 0.06 0.01 0.11

3.19 3.73 0.28 0.11 0.18 0.17 0.09 0.04 0.09

3.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

22.00 26.00 1.00 0.41 0.90 0.73 0.46 0.22 0.40

25.48 0.54 0.05 0.20 0.34 0.19 0.02 0.02 0.03 10.57

1.87 0.08 0.03 0.40 0.47 0.20 0.06 0.09 0.03 0.53

21.65 0.10 0.12 0.00 0.00 0.00 0.00 0.00 0.00 9.13

30.27 0.74 0.31 1.00 1.00 1.00 0.20 0.90 0.13 11.72

variables 327 320 327 328 297 317 323 319 326

Panel A3: Other control variables BankSize 359 LoanRatio 355 CapitalRatio 357 Listed 359 ForStgInvestor 359 PLGState 341 PLGForeign 341 PLGPrivate 341 ShareOth 334 GeoIndex 319

This table reports summary statistics on key variables. The sample is an unbalanced panel covering 359 bank-years over the period of 2003–2010. Panel A1 reports the summary statistics of bank performance variables. Panel A2 reports the summary statistics of board characteristics variables. Panel A3 reports the summary statistics of other control variables. See Table 1 for variable deﬁnitions.

addition, in our sample, about 20% of the banks are listed in stock exchanges, 34% of the banks have foreign strategic investors with the average foreign ownership 6%. The average controlling state ownership is 19%. We further divide the sample based on bank type and report the descriptive statistics in Table 2b Panel B. We ﬁnd that the boards in SOCBs and JSCBs are larger, meet more often, have more independent directors, politically-connected directors, busy directors and foreign directors than city commercial banks (CCBs) and rural commercial banks (RCBs). The SOCBs and RCBs separate the roles of the chairman of the board and bank president while the JSCBs and CCBs exhibit the CEO-chairman duality.16 Fig. 1 shows the time trends of board size, board meetings, board independence and the proportion of politically-connected directors from 2005 to 2010.17 Panel A shows that the mean board size remains relatively ﬂat for both the SOCBs and JSCBs while it has increased slightly for the CCBs and RCBs. The average board size is 12.76 in 2005 and 14.25 in 2010. Panel B reports the time trend in the number of board meetings. The CCBs and RCBs show a signiﬁcant upward trend while the SOCBs exhibit a swing over the sample

16 In addition, we test the statistical signiﬁcance of the differences in each board characteristic between (1) SOCBs and JSCBs, (2) SOCBs and CCBs, (3) JSCBs and CCBs, (4) CCBs and RCBs. The means of board Meetings are signiﬁcantly different across all groups. The means of BoardSize are signiﬁcantly different for groups (2) and (3). The means of Duality and IndepDirector are signiﬁcantly different for groups (3) and (4). The means of PoliticalDirector are signiﬁcantly different for group (4). To save space, the test statistics are not reported in the table. 17 The sample size is smaller in the beginning of the sample period (2003 and 2004). We restrict our trend analysis to the years after 2004 to ensure that the results are not driven by changes in sample composition.

2960

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

Table 2b Panel B: Board characteristics by bank type. SOCB

BoardSize Meetings Duality IndepDirector PoliticalDirector BusyDirector ForeignDirector OldDirector FemaleDirector

JSCB

CCB

RCB

N

Mean

Median

N

Mean

Median

N

Mean

Median

N

Mean

Median

26 26 26 26 26 25 26 26 26

15.85 10.23 0.00 0.28 0.48 0.14 0.14 0.02 0.15

16.00 9.50 0.00 0.27 0.42 0.13 0.12 0.00 0.13

84 83 84 84 75 80 83 84 84

15.79 8.59 0.05 0.29 0.37 0.20 0.1 0.01 0.09

15.00 8.00 0.00 0.33 0.37 0.15 0.07 0.00 0.08

182 178 183 183 168 178 180 177 182

12.74 6.08 0.13 0.19 0.36 0.10 0.04 0.01 0.11

13.00 5.00 0.00 0.18 0.38 0.00 0.00 0.00 0.09

35 33 34 35 28 34 34 32 34

13.00 5.48 0.00 0.22 0.33 0.11 0.01 0.02 0.09

13.00 5.00 0.00 0.20 0.30 0.00 0.00 0.00 0.07

The table reports summary statistics for board structure across state-owned commercial banks (SOCBs), joint-stock commercial banks (JSCBs), city commercial banks (CCBs) and rural commercial banks (RCBs). See Table 1 for variable deﬁnitions.

Panel A. Board Size

Panel C. % Independent Directors

Panel B. # of Board meetings

Panel D. % Directors who are politically connected

Fig. 1. Chinese bank board characteristics trends: 2005–2010. Fig. 1 shows the time trends of key board characteristics from 2005 to 2010. Panel A reports the average board size. Panel B reports the average number of board meetings. Panel C reports the percent of independent directors on the board. Panel D reports the percent of directors who are politically connected on the board. In addition, the percentage of woman directors on the board increases over this period. (To save space, we do not include the ﬁgures of other board variables.) See Table 1 for variable deﬁnitions.

period. On average, the number of meetings has increased from 5.08 on 2005 to 8.47 on 2010. Panel C shows that the mean percentage of independent directors generally increases over the sample period. The increase is most remarkable for the RCBs, followed by the CCBs and SOCBs. The percentage increases from about 19% in 2005 to about 26% in 2010. Panel D shows that the trend in the mean percentage of politically-connected directors differs across different types of banks. The SOCBs exhibit the biggest decline over the sample period, which is expected given the dramatic transformation process that they have gone through. The percentage of politicallyconnected directors remains relatively ﬂat for the JSCBs and CCBs until 2008, and then decreases slightly. The percentage of politically-connected directors exhibits a slight increase for the RCBs until 2009. We also ﬁnd that there is no strong time trend in the duality of CEO-Chairman duality, the percentage of old directors, the percentage of female directors. The mean percentages of busy directors and foreign directors increase slightly over the sample period.18

18

For brevity, we do not report all the ﬁgures. They are available upon request.

5. Empirical results 5.1. Basic regression results The main objective of this paper is to study empirically the impact of various board characteristics on bank performance and asset quality. We use OLS regressions with robust standard errors clustered at the bank level and include year dummies to account for the common shocks in the market or regulatory environment in a given year.19 We regress each performance variable on board variables, starting with four widely used board variables in the literature (namely board size, meeting, duality and independent directors) and then adding additional director characteristics in the separate regressions (including the percentage of politically-connected directors, busy directors, foreign directors, old directors and female directors). For the sake of space, we do not report all the tables. We

19 With a sample of 52 banks from the period of 2003–2010, we do not have enough observations to construct 2-way cluster-robust standard errors to account for both time-series and cross-sectional dependences. Hence, we run the OLS regression with standard errors clustered at the bank level and additional year dummies to control for the common shocks in a given year.

2961

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968 Table 3 Board characteristics and bank performance (ROA). ROA Intercept Board characteristics BoardSize Meetings Duality IndepDirector

ROA *

ROA

ROA

ROA

ROA

0.011 (1.16)

0.016 (1.79)

0.011 (1.15)

0.01 (1.05)

0.008 (0.87)

0.01 (1.03)

0.0003*** (2.85) 0.0001* (1.94) 0.0004 (0.50) 0.005* (1.82)

0.0003*** (2.59) 0.0001 (1.56) 0.001 (1.33) 0.004* (1.65) 0.004** (2.5)

0.0003*** (2.8) 0.0001* (1.82) 0.0004 (0.46) 0.003* (1.88)

0.0003*** (2.87) 0.0001* (2.00) 0.0001 (0.43) 0.005** (2.02)

0.0003*** (3.2) 0.0001* (1.93) 0.0001 (0.72) 0.005* (1.93)

0.0003*** (2.85) 0.0001* (1.95) 0.0001 (0.51) 0.005* (1.78)

PoliticalDirector BusyDirector

0.001 (0.52)

FoeignDirector

0.007 (1.62)

OldDirector

0.009 (1.34)

FemaleDirector Bank characteristics BankSize LoanRatio CapitalRatio Listed ForStgInvestor Ownership structure PLGState PLGForeign PLGPrivate ShareOth Geo-economy effect GeoIndex Year dummies N Adj-R2

0.001 (0.4) 0.0000 (0.68) 0.0001 (0.03) 0.065*** (4.99) 0.0001 (0.49) 0.001 (1.07)

0.0000 (0.61) 0.001 (0.3) 0.055*** (4.35) 0.0001 (0.26) 0.001 (1.19)

0.0000 (0.54) 0.002 (0.54) 0.059*** (4.47) 0.0001 (0.53) 0.001 (0.94)

0.0000 (0.9) 0.0001 (0.02) 0.062*** (4.76) 0.001 (0.61) 0.001 (1.58)

0.0000 (1.2) 0.002 (0.72) 0.06*** (4.6) 0.001 (0.62) 0.0001 (0.58)

0.0000 (0.7) 0.0001 (0.03) 0.065*** (4.99) 0.0001 (0.45) 0.001 (1.12)

0.005** (2.93) 0.014** (3.09) 0.002 (0.7) 0.005 (0.57)

0.003* (1.8) 0.012** (2.79) 0.003 (1.21) 0.004 (0.41)

0.005** (2.66) 0.015** (3.18) 0.002 (0.75) 0.002 (0.27)

0.005** (2.51) 0.008 (1.43) 0.001 (0.42) 0.005 (0.62)

0.006** (3.28) 0.014** (3.06) 0.002 (1) 0.006 (0.68)

0.005** (2.95) 0.014** (3.09) 0.002 (0.74) 0.004 (0.5)

0.001 (0.96) Yes 279 0.27

0.001 (1.49) Yes 267 0.31

0.0001 (0.65) Yes 271 0.27

0.001 (1.04) Yes 276 0.27

0.0001 (0.72) Yes 274 0.28

0.001 (0.81) Yes 279 0.27

The table presents the results of regressing ROA on various board characteristics variables. The table reports the regression coefﬁcients, t-statistics (in parentheses), number of observations (N) and adjusted-R2. The deﬁnitions of variables are in Table 1. * Statistical signiﬁcance at 10% level. ** Statistical signiﬁcance at 5% level. *** Statistical signiﬁcance at 1% level.

report the main empirical results, including the results of regressing ROA on all board variables in Table 320 and the results of regressing alternative performance and asset quality variables on ﬁve key board variables in Table 4. The unreported tables show similar results and are available upon request. Table 3 shows that board size has a signiﬁcantly negative relationship with ROA at the 1% level across all the models, which is consistent with a number of empirical studies (e.g. Hermalin and Weisbach, 2003; Yermack, 1996). The strong negative relationship supports the hypothesis that large boards represent inefﬁcient governance and have detrimental effects on bank performance in China. The coordination, communication, and decision making problems associated with large boards might outweigh the beneﬁt 20 ROA is a more comprehensive measure of proﬁtability and it is widely used in the literature, which allows comparison with previous studies on China or other countries.

of providing collective information, which eventually hinders the advisory and monitoring functions of the boards. The number of board meetings has shown a signiﬁcantly positive relationship with ROA at the 10% level, which is consistent with Andres and Vallelado (2008). The results support the hypothesis that frequent board meetings signal increased supervision of the top bank management in China. Board members might play a proactive role in the meetings to discuss and exchange ideas on how to monitor and advise managers, which could subsequently enhance bank performance. We observe similar relationships between the proportion of independent directors and performance of Chinese banks. The coefﬁcients on the proportion of independent directors are positive and signiﬁcant at the 10% level, which is consistent with previous banking studies using the samples in developed countries (e.g. Cornett et al., 2009; Andres and Vallelado, 2008). The results support the hypothesis that independent director alleviate conﬂicts of

2962

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

Table 4 Board characteristics and alternative measures of bank performance. ROE Intercept Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector Bank characteristics BankSize LoanRatio CapitalRatio Listed ForStgInvestor Ownership structure PLGState PLGForeign PLGPrivate ShareOth Geo-economy effect GeoIndex Year dummies N Adj-R2

Pre-provision proﬁt ***

**

NPL ratio **

Stock of NPLs **

NCO ratio

Level of NCOs

0.476 (3.01)

0.03 (2.25)

0.195 (2.06)

6.413 (3.55)

0.001 (0.05)

12.56** (2.77)

0.003* (1.95) 0.001 (0.97) 0.029* (1.95) 0.068* (1.67) 0.067** (2.61)

0.0002 (1.2) 0.0002** (2.08) 0.001 (1.03) 0.001 (0.27) 0.003* (1.67)

0.0002 (0.18) 0.001 (1.61) 0.005 (0.61) 0.016 (0.60) 0.01 (0.63)

0.006 (0.32) 0.030** (2.17) 0.213 (1.26) 1.255** (2.55) 0.659** (2.22)

0.0003 (1.22) 0.0002 (0.27) 0.002 (0.8) 0.013* (1.88) 0.005 (1.14)

0.007 (0.14) 0.046 (1.39) 0.149 (0.37) 2.556** (2.05) 1.04 (1.46)

0.003 (0.61) 0.04 (0.72) 1.18*** (5.39) 0.009 (0.67) 0.018* (1.69)

0.0001 (0.67) 0.014*** (3.01) 0.055*** (3.02) 0.0001 (0.34) 0.002** (2.04)

0.001 (0.29) 0.006 (0.19) 0.249* (1.9) 0.001 (0.12) 0.005 (0.77)

1.082*** (21.98) 2.116*** (3.36) 2.409 (0.96) 0.088 (0.57) 0.072 (0.59)

0.0001 (0.64) 0.006 (0.64) 0.003 (0.09) 0.001 (0.39) 0.001 (0.54)

1.174*** (9.65) 2.873* (1.87) 5.857 (0.92) 0.137 (0.37) 0.382 (1.27)

0.053* (1.68) 0.17** (2.23) 0.018 (0.44) 0.065 (0.43)

0.002 (0.81) 0.009 (1.37) 0.0001 (0.02) 0.0001 (0.01)

0.012 (0.66) 0.062 (1.36) 0.007 (0.28) 0.202** (2.24)

0.033 (0.09) 2.156** (2.48) 0.747 (1.63) 5.404** (3.08)

0.005 (0.9) 0.01 (0.8) 0.002 (0.27) 0.016 (0.65)

0.703 (0.8) 0.341 (0.16) 0.581 (0.53) 5.458 (1.32)

0.021** (2.07) Yes 267 0.20

0.002 (1.25) Yes 261 0.24

0.015** (2.46) Yes 264 0.11

0.146 (1.25) Yes 264 0.88

0.0001 (0.1) Yes 231 0.01

0.136 (0.47) Yes 231 0.64

The table presents the results of regressing alternative bank performance and asset quality variables on key board variables – BoardSize, Meetings, Duality, IndepDirector and PoliticalDirector. The table reports the regression coefﬁcients, t-statistics (in parentheses), number of observations (N) and adjusted-R2. The deﬁnitions of variables are in Table 1. * Statistical signiﬁcance at 10% level. ** Statistical signiﬁcance at 5% level. *** Statistical signiﬁcance at 1% level.

interests between insiders and shareholders and increase the efﬁcacy of supervision of top management. We ﬁnd new evidence that the proportion of politicallyconnected directors has a negative relationship with ROA at the 5% level, which is consistent with a stream of literature that uses non-banking samples (e.g. Fan et al., 2007; Boubakri et al., 2008). The negative effect suggests that politically connected boards might allow more government intervention and do not have incentives to maximize value and improve overall bank performance. Furthermore, the effect of board characteristics on bank performance appears to be, not only statistically signiﬁcant, but also economically large. Board size has the most economic signiﬁcance. An increase in board size by one standard deviation is associated with a decrease in ROA by 12.73%.21 The economic magnitude of the effect of the politically-connected directors is also quite signiﬁcant. An increase in the proportion of politically-connected directors by one

standard deviation is associated with a decrease in ROA by 9.57%. In addition, as the degree of board independence (board meeting frequency) rises by one standard deviation, ROA rises by 7.31% (4.96%). The adjusted R2 reported in Table 3 column (1) shows that approximately 27% of the variations in the ROA of banks in China can be explained when we include four board variables that have been extensively studied by the previous governance literature, namely board size, meeting frequency, duality and the degree of independence. By adding the proportion of politically-connected directors, the adjusted R2 increases by roughly 4–31%, which indicates the signiﬁcant role of political connection in the Chinese banking sector. Table 3 columns (3–5) show that there are no signiﬁcant relationships between ROA and other director characteristics variables. The adjusted R2 also shows that including other director characteristics does not increase the explanatory power.22

21 We multiply the standard deviation of board size by the coefﬁcient of board size in Table 3 (3.19 0.0003 = 0.00096). The result represents how much ROA changes, given that board size changes by one standard deviation. The average ROA in the sample is 0.00752. Therefore, a change in ROA of 0.00096 represents a percentage change of 12.73% of the average ROA (0.00096/0.00752 = 12.73%). The economic signiﬁcance of other board variables is calculated in the same manner.

22 We do not ﬁnd signiﬁcant impacts of duality on ROA, which is in line with Brickley et al. (1997). Such evidence is in support of the hypothesis that the dual position of the president and chairman of the board does not necessarily destroy ﬁrm value. One possible reason is that the dual position might reduce conﬂict between two powerful positions in decision-making. However, it is still important to include the duality variable in the empirical tests as an additional control.

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

For the control variables, we ﬁnd that the capital ratio has a signiﬁcantly positive impact on ROA, which indicates that banks with higher degree of capitalization perform better. Furthermore, we ﬁnd that the coefﬁcients on PLGState and PLGForeign are signiﬁcantly negative, which suggests that the percentage of shares owned by the controlling shareholder with the controlling shareholder being the state or foreign investor has a signiﬁcantly negative impact on bank performance. The results suggest that the controlling state ownership is negatively correlated with bank performance, which is consistent with the previous banking literature (e.g. Berger et al., 2009; Jia, 2009; Lin and Zhang, 2009). In addition, we ﬁnd new evidence that the controlling ownership of the foreign investor is negatively correlated with bank performance. Anecdotal evidence shows that foreign banks in China have signiﬁcantly declined in recent years.23 A few possible reasons include limited business scopes, restrictions on the scale of loans and insufﬁcient knowledge of business culture. The recent ﬁnancial crisis might also have contributed to the decline in foreign banks. Nevertheless, it is beyond the scope of our paper to explore in detail what could cause the decline of foreign banks. Furthermore, the negative correlation between ownership concentration and bank performance suggests the expropriation effects of controlling shareholders. A higher level of controlling ownership corresponds to weaker governance and a higher tendency to expropriate minority shareholders. Table 4 reports the results of regressing the alternative performance variables and additional asset quality variables on ﬁve key board variables which have more explanatory power as shown in Table 3, including board size, meeting, duality, the proportion of independent and politically-connected directors. The two alternative performance variables, ROE and pre-provision proﬁt, produce similar results. The ROE is positively correlated with the degree of political connection (at the 5% level) and negatively correlated with board size and the degree of board independence (at the 10% level). The pre-provision proﬁt is positively correlated with meeting frequency (at the 5% level) and negatively correlated with the degree of political connection (at the 10% level). A new ﬁnding is the signiﬁcantly negative relationship between duality and ROE. For the asset quality variables, we ﬁnd that the stock of NPLs is positively correlated with the degree of political connection (at the 5% level) and negatively correlated with meeting frequency and the degree of board independence (at the 5% level). The degree of board independence is negatively correlated with the NCO ratio and the level of NCOs.24 Table 4 also shows that the geo-economic index has a signiﬁcantly positive relationship with ROE and a signiﬁcantly negative relationship with the NPL ratio. Banks that operate in the more developed areas have higher ROE and the lower amount of NPLs, which provide support for the location effect. Foreign strategic investors have a signiﬁcantly positive impact on ROE and pre-provision proﬁt. In addition, the ShareOth variable is negatively correlated with the NPL ratio and the stock of NPLs, which suggests that banks with more diverse ownership have better asset quality. Overall, the OLS results are in support of the hypothesis that the board of directors plays a signiﬁcantly important role in Chinese banks. We ﬁnd that board meeting frequency and the proportion of independent directors have signiﬁcantly positive impacts on bank performance and asset quality while the proportion of

23 According to the International Financial Market Report released by the People’s Bank of China in April, 2010, the market share of foreign banks in China fell from 2.16% in late 2008 to 1.71% by the end of 2009. 24 It is hard to accurately measure the asset quality of the Chinese banks given their dramatic restructuring program through recapitalization and NPLs carve-out. GarcíaHerrero et al. (2009) point out that the stock of NPLs and NPL ratio are the preferred proxies for asset quality of banks in a transitional state.

2963

politically-connected director has a signiﬁcantly negative impact. We also document a signiﬁcantly negative impact of board size on bank performance. 5.2. Results for regional banks The regional banks, including the CCBs and RCBs, have strong local focuses. Although they have been gradually transformed into joint-stock banks, the local government is often the largest shareholder. The regional banks mostly handle locally developed projects and programs. Ferri (2009) ﬁnds evidence that the performance of CCBs is related to the banks’ locations. To further examine the location effect and whether it dominates the impact of boards in the regional banks, we re-estimate the main regressions using the sub-sample of CCBs and RCBs. The results are reported in Table 5. For brevity, we present the empirical results with ROA as the dependent variable. The ﬁrst column of Table 5 shows the results with the geo-economic index as the control variable for the location effect. We ﬁnd that for the regional banks, meeting frequency and the proportion of independent directors are positively correlated with ROA at the 1% and 5% level respectively. The geo-economic index is positively correlated with ROA at the 10% level. Furthermore, following Ferri (2009), we classify the regions based on the level of economic development. We assign dummy variables for the CCBs and RCBs located in different regions. Rich equals one if the bank’s geo-economic index is in the top 25th percentile and zero if not. MRich equals one if the bank’s geo-economic index is in the middle 50th percentile and zero otherwise. We re-estimate the regressions with two regional dummies. The results are reported in the second column of Table 5. The positive impacts of meeting frequency and the proportion of independent directors on ROA remain. The two location dummy variables, Rich and MRich, are positively correlated with ROA at the 10% and 1% level respectively. Overall, we ﬁnd strong evidence to support the location effect in the regional banks, which is consistent with Ferri (2009). The regional banks that are located in the more developed areas have better performance and asset quality. In addition, we ﬁnd that among various board characteristics, board meeting frequency and the proportion of independent directors play more important roles in the regional banks. 5.3. Results for banks with different characteristics The impact of board characteristics on bank performance could be shaped by the regulatory frameworks. We examine the differential effects of board characteristics on performance across the banks that might be subject to different regulatory requirements for interest rates and capital adequacy.25 5.3.1. Results for large and small banks We expect that the impact of the regulatory changes would differ across the banks with different sizes. The gradual deregulation 25 Regulation governing the interest rate: After a series of reforms, China has made substantial progress in liberalizing its interest rates. In the retail lending and deposit markets, the deposit-rate ﬂoor and the lending-rate ceiling were eliminated in October 2004, except in respect of credit cooperatives. The current interest rate system is a dual-track system with a deposit-rate ceiling and a lending-rate ﬂoor imposed by the People’s Bank of China (PBC). The average spread is about 3%. The lending-rate ﬂoor for all commercial banks was set to be 90% of the benchmark lending rate. Regulation governing the capital adequacy ratio: During our sample period of 2003–2010, there were two regulatory changes in the capital adequacy requirement. In February 2004, the CBRC issued the new rule that required all commercial banks to maintain a minimum capital adequacy ratio (CAR) of 8% and set January 1, 2007 as the deadline for the ﬁnal compliance. In December of 2009, the CBRC lifted the required minimum CAR to 11% for large and systematically important banks and 10% for small to medium size banks.

2964

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

Table 5 Board characteristics and bank performance: Regional banks. Regional banks

Intercept Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector Bank characteristics BankSize LoanRatio CapitalRatio Listed ForStgInvestor Ownership structure PLGState PLGForeign PLGPrivate ShareOth Geo-economy effect GeoIndex

ROA

ROA

0.029** (2.29)

0.043*** (3.35)

0.0001 (0.59) 0.0003*** (3.04) 0.001 (0.84) 0.005* (1.67) 0.001 (0.62)

0.0001 (0.64) 0.0003** (2.55) 0.001 (0.46) 0.005* (1.65) 0.001 (0.47)

0.0002** (2.84) 0.001 (0.16) 0.004 (0.25) 0.001 (0.48) 0.001* (1.66)

0.0002** (2.75) 0.001 (0.31) 0.01 (0.61) 0.0001 (0.11) 0.002* (1.8)

0.017** (5.08) 0.02*** (3.34) 0.003 (1.24) 0.028 (1.35)

0.014** (4.16) 0.017** (2.88) 0.004 (1.5) 0.019 (1.23)

0.002* (1.95)

Rich MRich Year dummies N Adj-R2

Yes 172 0.38

0.001* (1.72) 0.002*** (2.65) Yes 172 0.39

The table presents the robustness test results of regressing ROA on board characteristics variables - BoardSize, Meetings, Duality, IndepDirector and PoliticalDirector for regional banks, namely city commercial banks (CCBs) and rural commercial banks (RCBs). The table reports the regression coefﬁcients, t-statistics (in parentheses), number of observations (N) and adjusted-R2. The deﬁnitions of variables are in Table 1. * Statistical signiﬁcance at 10% level. ** Statistical signiﬁcance at 5% level. *** Statistical signiﬁcance at 1% level.

of interest rates is viewed as one of the critical regulatory efforts to break China’s banking monopoly. It is expected to change large banks’ business scopes and provide more competitive advantages to small and medium banks. In addition, large banks face stricter capital adequacy requirements and are subject to more scrutiny. The regulators pay particular attention to large banks that might pose a ‘‘systematic risk’’. We divide the sample into two size groups.26 To highlight the differences between the large banks and other banks, we deﬁne large banks in a given year to be banks whose total assets are above the

26 While our multivariate estimations include ﬁrm size as a control variable, which essentially controls for the level of bank performance across ﬁrm size, it may be that the impacts of board characteristics on bank performance vary across ﬁrm size.

25th percentile of the sample banks that year. Otherwise the banks are classiﬁed as small. The sub-sample regression result is reported in Table 6. We ﬁnd that for large banks, board size, duality and the proportion of politically-connected directors are negatively correlated with ROA at the 10% level. Meeting frequency and the proportion of independent directors are positively correlated with ROA at the 5% and 10% level respectively. For small banks, meeting frequency and the proportion of independent directors are positively correlated with ROA at the 1% and 10% level respectively. In addition, the explanatory power of the model is higher for large banks. 5.3.2. Results for banks with high and low capital adequacy ratios We also expect that banks with different capital adequacy ratios behave differently since they face different pressures in meeting the new capital adequacy requirements. Chiuri et al. (2002) ﬁnd that the enforcement of capital adequacy requirements has a negative effect on the supply of bank loans and the effect tends to be stronger for initially less capitalized banks. The degree of a bank’s capitalization reﬂects the magnitude of regulatory pressure experienced by the bank. We divide the sample into two groups based on the bank’s capital adequacy ratio (CAR).27 We classify the banks with CAR lower than the median level as low capitalized and those with CAR above the median level as well capitalized. Table 6 reports the sub-sample regression results. We ﬁnd that for the well capitalized banks, board duality and the proportion of politically-connected directors are negatively correlated with ROA at the 5% level. The proportion of independent directors is positively correlated with ROA at the 1% level. For the low capitalized banks, board size is negatively correlated with ROA at the 1% level and the meeting frequency is positively correlated with ROA at the 5% level. Overall, the results show that the impact of board characteristics on performance differs between banks with different characteristics. The board of directors plays a stronger role in large banks and banks with low capital adequacy ratios. The results complement research in Coles et al. (2008) and Raheja (2005), which suggest that optimal board structure depends on bank characteristics. However, the results for this analysis should be cautiously interpreted due to low statistical power from the small sample size for each group. 5.4. Additional Robustness checks We conduct additional robustness checks but do not report them in tables for the sake of space. Since there might be lagged impacts of the board characteristics on bank performance and asset quality, we re-estimate the main regression using the board variables that are lagged by 1 year. The results remain similar. We specify alternative dependent variables. We recalculate the NPL ratio and NCO ratio as the percentage of total assets instead of total loan. We ﬁnd similar results. In addition, we include an additional dependent variable, the loan loss provision ratio, which is measured as the amount of loan loss provision divided by total loans. We do not ﬁnd signiﬁcant impacts of board structure on the loan loss provision ratio. Bank size and the proportion of independent directors might have non-linear relationships with performance and asset quality. Following the previous literature (e.g. Andres and Vallelado, 2008), we include an additional quadratic term of bank size or a quadratic term of the proportion of independent directors in the regressions.

27 We include the capital adequacy ratio as an additional control variable in the regressions. The main results are similar. We do not ﬁnd a signiﬁcant relationship between the capital adequacy ratio and bank performance.

2965

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968 Table 6 Board characteristics and bank performance: Banks with different characteristics. ROA Sorted by bank size Large Intercept Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector Bank characteristics BankSize LoanRatio CapitalRatio List ForStgInvestor Ownership structure PLGState PLGForeign PLGPrivate ShareOth Geo-economy effect GeoIndex Year dummies N Adj-R2

ROA Sorted by capital adequacy ratio Small

Well capitalized

Low capitalized

0.088** (3.38)

0.035*** (3.06)

0.03*** (2.79)

0.0004** (2.65) 0.0001** (2.67) 0.004** (2.28) 0.007* (1.70) 0.004** (2.74)

0.0001 (0.89) 0.0003*** (2.80) 0.001 (0.57) 0.005* (1.84) 0.0004 (0.21)

0.0001 (0.85) 0.0001 (0.89) 0.003** (2.66) 0.01*** (3.40) 0.004** (2.13)

0.0004*** (2.84) 0.0002** (2.22) 0.001 (0.47) 0.002 (0.43) 0.001 (0.71)

0.003*** (5.39) 0.0001 (0.03) 0.002 (0.07) 0.001 (0.62) 0.001* (1.84)

0.002** (3.42) 0.008** (2.12) 0.032** (2.04) 0.001 (1.14) 0.002** (2.59)

0.0001 (0.85) 0.018** (4.28) 0.039*** (2.68) 0.0001 (0.09) 0.0001 (0.17)

0.001 (1.58) 0.019*** (3.47) 0.101*** (3.53) 0.002* (1.96) 0.002** (2.57)

0.003 (1.39) 0.016** (2.57) 0.009** (3.2) 0.018** (2.1)

0.017** (5.37) 0.021** (3.7) 0.002 (0.51) 0.02 (1.51)

0.0001 (0.02) 0.011* (1.86) 0.001 (0.24) 0.004 (0.38)

0.006** (2.4) 0.019** (2.88) 0.002 (0.47) 0.004 (0.26)

0.001 (1.15) Yes 76 0.52

0.001* (1.93) Yes 191 0.41

0.001 (1.00) Yes 136 0.35

0.001 (1.16) Yes 129 0.44

0.004 (0.28)

The table presents the robustness test results of regressing ROA on board characteristics variables for banks with different sizes and different capital adequacy ratios. Large banks are deﬁned as the banks whose total assets are above the 25th percentile of the sample banks in a given year. Low capitalized banks are deﬁned as the banks with the capital adequacy ratio lower than the median level. The table reports the regression coefﬁcients, t-statistics (in parentheses), number of observations (N) and adjusted-R2. The deﬁnitions of variables are in Table 1. * Statistical signiﬁcance at 10% level. ** Statistical signiﬁcance at 5% level. *** Statistical signiﬁcance at 1% level.

We do not ﬁnd a non-linear relationship and our main results still hold. We use different measures to control for ownership structure. The additional ownership variables are: the percentage of state, foreign and private ownership; the dummy variables indicating whether the largest shareholder is state, private or foreign. The main results are similar. The percentage of state ownership has a signiﬁcantly negative impact on bank performance. Banks with the largest shareholder being the state, state-owned enterprise or foreign investor have poorer performance. In addition, we include a squared term for each of the ownership variables and examine possible non-linear relationships between ownership concentration and bank performance. We do not ﬁnd evidence for the nonlinear effect. The results of all robustness tests are available upon request. We also run robustness tests using an alternative geoeconomic index constructed with the GDP per capita of the provinces that a bank has a branch in. The results are similar.

5.5. Discussion of endogeneity 5.5.1. GMM estimation A key concern for any analysis of board effects is the endogeneity of board structure, a point made both theoretically and empirically by Hermalin and Weisbach (1998, 2003), among others. We argue that endogeneity might not be a serious problem for our study. Boards are newly developed in the Chinese banking sector. They are generally put into place by the Chinese regulatory authority as one of the reform efforts and the board structures of most banks are set up following ‘‘best practices’’ in the Anglo-Saxon system. In other words, bank boards in China are structured based on guidelines and requirements imposed from outside the banks. Such external inﬂuences are thus likely exogenous. In addition, the frequency of board turnover is low. A modern director nomination and election mechanism is not in place. The high costs of termination and search costs also impose limitations on the adjustment of

2966

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

board composition. Hence, bank boards in China are likely exogenously determined. Nevertheless, we take measures to control for the potential endogeneity problems. Following Andres and Vallelado (2008), we employ the generalized method of moments (GMMs) and use the two-step system estimator with adjusted standard error for potential heteroskedasticity proposed by Arellano and Bond (1998). The approach accounts for the unobserved heterogeneity and the dynamic nature of panel data. We use lagged board variables and lagged state ownership as instruments. The logic is that board and ownership variables in earlier years could not have resulted from bank performance in subsequent years. Therefore, endogeneity is unlikely. Since out sample size is not large, we use the adjustment for small sample as in Windmeijer (2000). We report the system estimator regression results in Table A1. The results show that board size and the proportion of politically-connected directors have signiﬁcantly negative impacts on the ROA and ROE the 10% level. The meeting frequency is negatively correlated with the stock of NPLs at the 10% level. The proportion of politically-connected directors is positively correlated with the stock of NPLs at the 10% level. To test the model speciﬁcation validity, we calculate the Hansen/ Sargen test of over-identiﬁcation of restrictions and the F test of joint signiﬁcance for all independent variables. The F test statistics are signiﬁcant but quite low. The Hansen tests show no rejection of the null hypothesis and conﬁrm the validity of the choices of instrument variables. Furthermore, we perform the endogeneity tests to empirically test the endogeneity of board variables. The endogeneity tests cannot reject the null hypothesis that the board variables are strictly exogenous (the p-values are all above 10%). If the instrument is not too weak as shown in the Hansen tests, the insigniﬁcant endogeneity tests suggest that OLS estimators are unbiased, Overall, the system estimators still point to the relevance of analyzing board characteristics in the Chinese banking sector. We ﬁnd similar impacts of board size, meeting frequency and the proportion of politically-connected directors on bank performance and asset quality. Furthermore, the endogeneity test results support our conjecture that the bank board in China is exogenously determined and conﬁrm the validity of the OLS results. 5.5.2. Additional tests of reverse causality We run additional robustness tests to address the reverse causality issue. It is conceivable that banks with better performance may be chosen to ﬁrst adopt the modern governance mechanism. To test whether the direction of causality runs from bank performance to board structure, we explore whether bank performance inﬂuence the choice of board structure during recent ﬁnancial crisis. We focus on the sub-period around the ﬁnancial crisis because the ﬁnancial crisis of 2008 can be regarded as an exogenous shock on Chinese bank performance. It originated in the US and was unrelated to governance reforms in China. Speciﬁcally, we examine whether banks adjust their board structures in response to changes in performance during the ﬁnancial crisis. First, we divide the sample based on changes in bank performance around the ﬁnancial crisis period, measured as difference between ROA in 2009 and ROA in 2007. Banks with change in ROA being above the median level are placed in one group and those with change in ROA being below the median level in the other group. We then compare board variables, including board independence and board size, between the two groups. The results show that there are no signiﬁcant changes in board structure between banks with better performance and those with poorer performance around the ﬁnancial crisis period. Second, we regress board variables on a post-crisis dummy variable, which is equal to one for the post-crisis period (after 2008) and zero otherwise, and bank size. We ﬁnd that the coefﬁcient

on the post-crisis dummy variable is not signiﬁcant in all regressions. The results suggest that board variables, including board independence and board size, do not change signiﬁcantly in response to changes in performance before and after the crisis. In summary, the empirical results28 of robustness tests indicate that bank board structure is not inﬂuenced by bank performance and reverse causality does not appear to be a problem for our study. 6. Conclusions In the context of the ongoing bank reform in China, this paper examines the role of boards of directors in the Chinese banking sector. We use a panel data of the top 50 banks from 2003 to 2010, a recent period following dramatic changes in the Chinese banking sector, including the restructuring of the ‘‘big four’’ state-owned banks and the full opening of the banking sector to foreign competition. We study a comprehensive set of board characteristics (size, composition and functioning of the board) and analyze the impacts of board characteristics on bank performance and asset quality. There is strong evidence that board size and the proportion of politically-connected directors have signiﬁcantly negative impacts on bank performance and loan quality. We also ﬁnd evidence that the number of board meetings and the proportion of independent directors have signiﬁcantly positive impacts on bank performance and asset quality. The results suggest that in the case of Chinese banks, small boards, boards that meet more frequently and boards with more independent and less politicallyconnected directors tend to be more efﬁcient in supervising and advising functions. Overall, we ﬁnd that the board of directors plays a signiﬁcant role in Chinese bank governance, and certain characteristics of bank board affect the motivation and abilities of the board in its supervisory and advisory duties and subsequently impact bank performance and asset quality. The ﬁndings of this paper have important policy implications. In particular, it suggests that a sound corporate governance system becomes critical in the Chinese banking sector while the banks undergo a remarkable transformation into modern market-oriented ﬁnancial enterprises and start to face the agency problems similar to those in the Anglo-Saxon system. Although China’s bank reforms are still ongoing and thus it is hard to draw any conclusions on how it may eventually affect the functioning of the banking system, the knowledge of the current governance structure and how it impacts bank performance in this study helps shed light on the likely direction and effects of future reforms. Acknowledgements We especially would like to express our gratitude to Jian Yang who participated in an earlier version of the paper and provided continuous guidance and assistance in the preparation of this paper. We also thank an anonymous referee for helpful comments. Liang acknowledges the support from the National Social Science Foundation of China (09&ZD037) and National Natural Science Foundation of China (71172066). The project is also supported by the Fundamental Research Funds for the Central Universities (NKZXA1212). Part of this research was conducted while Jiraporn served as a Visiting Associate Professor at Thammasat University and The National Institute of Development Administration (NIDA) in Bangkok, Thailand. Appendix A See Table A1. 28

The results of all robustness tests are available upon request.

2967

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968 Table A1 Board characteristics and bank performance: System estimator. ROA Intercept Board characteristics BoardSize Meetings Duality IndepDirector PoliticalDirector Bank characteristics BankSize LoanRatio CapitalRatio List ForStgInvestor Ownership structure PLGState PLGForeign PLGPrivate ShareOth Geo-economy effect GeoIndex Year dummies F test Hansen test Endogeneity test

ROE

Pre-provision proﬁt **

**

NPL ratio

Stock of NPLs

NCO ratio

Level of NCOs

0.02 (0.20)

6.4 (2.72)

0.3 (0.72)

17.0** (2.84)

0.00008 (0.41) 0.00005 (0.30) 0.001 (0.55) 0.002 (0.37) 0.004 (1.12)

0.0007 (1.07) 0.00004 (0.05) 0.01 (0.77) 0.03 (1.09) 0.004 (0.37)

0.03 (1.08) 0.05* (1.65) 0.2 (0.61) 1.1 (1.33) 0.7* (1.78)

0.001 (0.57) 0.001 (0.46) 0.10 (0.82) 0.20 (0.86) 0.07 (0.86)

0.00003 (0.00) 0.10 (1.63) 0.10 (1.36) 0.20 (0.07) 0.5 (0.41)

0.005 (1.05) 0.01 (0.19) 1.2*** (4.84) 0.003 (0.28) 0.02 (1.25)

0.0004 (0.96) 0.01* (2.27) 0.04 (1.49) 0.002 (1.71) 0.001 (1.25)

0.0006 (0.26) 0.010 (0.35) 0.03 (0.26) 0.005 (0.72) 0.0007 (0.13)

1.1*** (17.35) 1.9* (2.18) 1.9 (0.59) 0.05 (0.31) 0.1 (0.77)

0.006 (0.80) 0.2 (0.78) 0.2 (0.65) 0.01 (0.67) 0.003 (0.27)

1.2*** (7.87) 4.1 (1.80) 7.6 (0.95) 0.5 (1.23) 0.5 (1.32)

0.004* (2.05) 0.01** (2.70) 0.003 (0.95) 0.006 (0.62)

0.08* (2.39) 0.2* (2.58) 0.03 (0.57) 0.03 (0.16)

0.0007 (0.27) 0.007 (1.33) 0.0004 (0.11) 0.003 (0.28)

0.01 (0.72) 0.05 (1.63) 0.004 (0.19) 0.05 (0.67)

0.1 (0.31) 2.0* (1.98) 1.1 (1.70) 5.5** (2.71)

0.003 (0.08) 0.1 (0.63) 0.05 (0.72) 0.4 (0.82)

-0.8 (0.76) 0.4 (0.15) 0.5 (0.30) 6.9 (1.41)

0.001 (1.42) Yes 7.75 [0.000] 0.04 [0.85] 5.01 [0.42]

0.03* (2.14) Yes 3.76 [0.000] 0.74 [0.39] 2.81 [0.73]

0.002* (2.35) Yes 4.66 [0.000] 0.02 [0.90] 4.52 [0.48]

0.005 (0.48) Yes 2.87 [0.000] 2.56 [0.11] 1.55 [0.91]

0.20 (1.19) Yes 21.03 [0.000] 5.95 [0.21] 3.97 [0.55]

0.01 (0.55) Yes 0.07 [1.00] 0.33 [0.57] 1.00 [0.96]

0.05 (0.14) Yes 29.24 [0.000] 0.61 [0.44] 8.08 [0.15]

0.02 (1.51)

0.5 (2.60)

0.04 (2.97)

0.0002* (1.65) 0.00004 (0.29) 0.001 (0.50) 0.0010 (0.24) 0.005* (2.27)

0.004* (2.08) 0.0008 (0.33) 0.03 (0.96) 0.07 (1.02) 0.07* (2.07)

0.0002 (0.75) 0.002 (0.46) 0.06** (2.93) 0.0002 (0.27) 0.0007 (1.13)

**

The table presents the results of the two-step system GMM estimate of regressing ROA on board characteristics variables. Variables in italics are instrumented through the GMM procedure. F test statistics is the test of model statistical signiﬁcance. Hansen test statistics is the test of over-identifying restrictions. Endogeneity test statistics is the test of endogenous regressors. The deﬁnitions of variables are in Table 1. Figures in parenthesis are t-statistics of regression coefﬁcients while p-values are in brackets. * Statistical signiﬁcance at 10% level. ** Statistical signiﬁcance at 5% level. *** Statistical signiﬁcance at 1% level.

References Adams, R.B., Hermalin, B., Ferreira, D., 2005. Powerful CEOs and their impact on corporate performance. Review of Financial Studies 18 (4), 1403–1432. Adams, R.B., Hermalin, B., Weisbach, M.S., 2010. The role of boards of directors in corporate governance: a conceptual framework and survey. Journal of Economic Literature 48 (1), 58–107. Adams, R.B., Ferreira, D., 2009. Women in the boardroom and their impact on governance and performance. Journal of Financial Economics 94, 291–309. Adams, R.B., Mehran, H., 2003. Is corporate governance different for bank holding companies? Federal Reserve Bank of New York Economic Policy Review, 123–142. Adams, R.B., Mehran, H., 2012. Bank board structure and performance: evidence for large bank holding companies. Journal of Financial Intermediation 21, 243–267. Agrawal, A., Knoeber, C., 1996. Firm performance and mechanisms to control agency problems between managers and shareholders. Journal of Financial and Quantitative Analysis 31 (3), 377–397. Ahn, S., Jiraporn, P., Kim, Y.S., 2010. Multiple directorships and acquirer returns. Journal of Banking and Finance 34, 2011–2026. Arellano, M., Bond, S., 1998. Dynamic Panel Data Estimation using DPD98 for GAUSS: A Guide for Users. Working Paper. Institute for Fiscal Studies. Andres, P., Vallelado, E., 2008. Corporate governance in banking: the role of the board of directors. Journal of Banking and Finance 32, 2570–2580.

Basel Committee on Banking Supervision, 2006. Enhancing Corporate Governance for Banking Organizations. Berger, A.N., Hasan, I., Zhou, M., 2009. Bank ownership and efﬁciency in China: what will happen in the world’s largest nation? Journal of Banking and Finance 33, 113–130. Berger, A.N., Hasan, I., Zhou, M., 2010. The effects of focus versus diversiﬁcation on bank performance: evidence from Chinese banks. Journal of Banking and Finance 34, 1417–1435. Booth, J.R., Cornett, M.M., Tehranian, H., 2002. Boards of directors, ownership, and regulation. Journal of Banking and Finance 26, 1973–1996. Boubakri, N., Cosset, J.C., Saffar, W., 2008. Politically connected newly privatized ﬁrms. Journal of Corporate Finance 14, 654–673. Brickley, J.A., Coles, J.L., Jarrell, G., 1997. Leadership structure: separating the CEO and chairman of the board. Journal of Corporate Finance 3, 189–220. Chen, G., Firth, M., Xu, L., 2009. Does the type of ownership control matter? Evidence from China’s listed companies. Journal of Banking and Finance 33, 171–181. Chiuri, M.C., Ferri, G., Majnoni, G., 2002. The macroeconomic impact of bank capital requirements in emerging economies: capital requirements in emerging economies: past evidence to assess the future. Journal of Banking and Finance 26, 881–904. Coles, J.L., Daniels, N.D., Naveen, L., 2008. Boards: does one size ﬁt all? Journal of Financial Economics 87, 329–356.

2968

Q. Liang et al. / Journal of Banking & Finance 37 (2013) 2953–2968

Core, J., Holthausen, R., Larcker, D., 1999. Corporate governance, chief executive ofﬁcer compensation, and ﬁrm performance. Journal of Financial Economics 51, 371–406. Cornett, M.M., McNutt, J.J., Tehranian, H., 2009. Corporate governance and earnings management at large US bank holding companies. Journal of Corporate Finance 15, 412–430. Dalton, D.R., Catherine, M.D., Jonathan, L.J., Alan, E.E., 1999. Number of directors and ﬁnancial performance: a meta-analysis. Academy of Management Journal 42, 674–686. Denis, D., McConnell, J., 2003. International corporate governance. Journal of Financial and Quantitative Analysis 38, 1–36. Eisenberg, T., Sundgren, S., Wells, M.T., 1998. Larger board size and decreasing ﬁrm value in small ﬁrms. Journal of Financial Economics 48, 35–54. Fama, E.F., Jensen, M.C., 1983. Separation of ownership and control. Journal of Law and Economics 26, 301–325. Fan, J.P.H., Wong, T.J., Zhang, T., 2007. Politically connected CEOs, corporate governance and post-IPO performance of China’s partially privatized ﬁrms. Journal of Financial Economics 84 (2), 330–357. Ferri, G., 2009. Are new tigers supplanting old mammoths in China’s banking system? Evidence from a sample city commercial banks. Journal of Banking and Finance 33, 131–140. Fich, E., Shivadasani, A., 2004. Are busy boards effective monitors? Journal of Finance 61, 689–724. Firth, M., Lin, C., Liu, P., Wong, S., 2009. Inside the black box: bank credit allocation in China’s private sector. Journal of Banking and Finance 33, 1144–1155. Fu, X., Heffernan, S., 2009. The effects of reform on China’s bank structure and performance. Journal of Banking and Finance 33, 39–52. García-Herrero, A., Gavilá, S., Santabárbara, D., 2009. What explains the low proﬁtability of Chinese banks? Journal of Banking and Finance 33, 2080–2092. Goyal, V.K., Park, C.W., 2002. Board leadership structure and CEO turnover. Journal of Corporate Finance 8, 49–66. Güner, A.B., Malmendier, U., Tate, G., 2008. Financial expertise of directors. Journal of Financial Economics 88, 323–354. Harris, M., Raviv, A., 2008. A theory of board control and size. Review of Financial Studies 21 (4), 1797–1832. Hermalin, B.E., Weisbach, M.S., 1998. Endogenously chosen boards of directors. American Economic Review 88 (1), 96–118. Hermalin, B.E., Weisbach, M.S., 2003. Boards of directors as an endogenously determined institution: a survey of the economic literature. Federal Reserve Bank of New York Economic Policy Review, 7–26. Jensen, M.C., 1993. The modern industrial revolution, exit, and the failure of internal control systems. Journal of Finance 48, 831–857. Jia, C., 2009. The effect of ownership on the prudential behavior of banks – the case of China. Journal of Banking and Finance 33, 77–87. Jirapon, P., Singh, M., Lee, C.I., 2009. Ineffective corporate governance: director busyness and board committee memberships. Journal of Banking and Finance 33, 819–828.

Klein, A., 2002. Audit committee, board of director characteristics, and earnings management. Journal of Accounting and Economics 35, 375–400. Laeven, L., Levine, R., 2009. Bank governance, regulation and risk taking. Journal of Financial Economics 93, 259–275. Levine, R., 2004. The Corporate Governance of Banks: A Concise Discussion of Concepts and Evidence. Working Paper. World Bank. Lin, X., Zhang, Y., 2009. Bank ownership reform and bank performance in China. Journal of Banking and Finance 33, 20–29. Macey, J.R., O’Hara, M., 2003. The corporate governance of banks. Federal Reserve Bank of New York Economic Policy Review, 91–107. Minton, B., Taillard, J., Williamson, R., 2011. Do Independence and Financial Expertise of the Board Matter for Risk Taking and Performance? Working Paper. The Ohio State University. Nguyen, B.D., Nielsen, K.M., 2010. The value of independent directors: evidence from sudden death. Journal of Financial Economics 98, 550–567. OECD, 2006. Policy Brief on Corporate Governance of Banks in Asia. Asian Roundtable on Corporate Governance. Oxelheim, L., Randoy, T., 2003. The impact of foreign board membership on ﬁrm value. Journal of Banking and Finance 27, 2369–2392. Pathan, S., 2009. Strong boards, CEO power and bank risk-taking. Journal of Banking and Finance 33, 1340–1350. Pi, L., Timme, S.G., 1993. Corporate control and bank efﬁciency. Journal of Banking and Finance 17, 515–530. Raheja, C.G., 2005. The determinants of board size and composition: a theory of corporate boards. Journal of Financial and Quantitative Analysis 40, 283–306. Rosenstein, S., Wyatt, J.G., 1990. Outside directors, board independence, and shareholder wealth. Journal of Financial Economics 26, 175–191. Rowe, W., Shi, W., Wang, C., 2011. Board Governance and Performance of Chinese Banks. Working Paper. University of Nebraska at Omaha. Tam, K.O., 1999. Models of corporate governance for Chinese companies. Corporate Governance: An International Review 8 (1), 52–64. Vafeas, N., 1999. Board meeting frequency and ﬁrm performance. Journal of Financial Economics 53, 113–142. Wang, C., 2005. Ownership and operating performance of Chinese IPOs. Journal of Banking and Finance 29, 1835–1856. Wei, Z., Xie, F., Zhang, S., 2005. Ownership structure and ﬁrm value in china’s privatized ﬁrms. Journal of Financial and Quantitative Analysis 40 (1), 87–108. Weisbach, M.S., 1988. Outside directors and CEO turnover. Journal of Financial Economics 20, 431–460. Windmeijer, F., 2000. A Finite Sample Correction for the Variance of Linear Twostep GMM Estimators. Working Paper. The Institute for Fiscal Studies. Wu, W., Wu, C., Rui, O., 2012. Ownership and the value of political connections: evidence from China. European Financial Management 18 (4), 695–729. Xu, N., Xu, X., Yuan, Q., forthcoming. Political connections, ﬁnancing friction, and corporate investment: evidence from Chinese listed family ﬁrms. European Financial Management. Yermack, D., 1996. Higher market valuation of companies with a small board of directors. Journal of Financial Economics 40, 185–211.

Board characteristics and Chinese bank performance

Board characteristics and Chinese bank performance

Recommend Documents