Managerial professional connections versus political connections: Evidence from firms' access to informal financing resources

Journal of Corporate Finance 41 (2016) 179–200

Contents lists available at ScienceDirect

Journal of Corporate Finance journal homepage: www.elsevier.com/locate/jcorpfin

Managerial professional connections versus political connections: Evidence from firms' access to informal financing resources☆ Qigui Liu a, Jinbo Luo b, Gary Gang Tian c,⁎ a b c

School of Management, Zhejiang University, Xihu District, Hangzhou, Zhejiang, 310058, China School of Accounting, Jiangxi University of Finance and Economics, China Department of Applied Finance and Actuarial Studies, Macquarie University, 2109, Australia

a r t i c l e

i n f o

Article history: Received 6 April 2016 Received in revised form 12 September 2016 Accepted 13 September 2016 Available online 16 September 2016 JEL classifications: G32 G34 Keywords: Trade credit Managerial professional connections Political connections Informal financing

a b s t r a c t This study investigates how managerial professional connections, through executives' membership of an industry association, play a role in helping firms obtain trade credit, while political connections do not. We document that firms whose managers have professional connections receive more trade credit, especially in firms that are not controlled by the state (non-SOEs), which have limited access to formal financial resources. The business environment, for example, low social trust and high product market competition, also strengthen the positive relationship between managers' professional connections and firms' access to trade credit. We further provide evidence that directors' professional connections also bring firms more trade credit and that firms with professional connections make more use of financing component of trade credit and abnormal trade credit. Our results are robust to a series of robustness and endogeneity tests. Overall, we argue that managerial professional connections, other than political connections, help firms, especially those with limited access to formal financing, to obtain informal financing resources. Crown Copyright © 2016 Published by Elsevier B.V. All rights reserved.

1. Introduction One central element of corporate finance is firms' access to financial resources. As a major channel of informal financing for firms, trade credit has been widely documented to play an important role around the world (Rajan and Zingales, 1995; Kohler et al., 2000; Guariglia and Mateut, 2006), and it is even more important in emerging markets like China, where the formal

☆ We are grateful for the valuable comments received from Nuttawat Visaltanachoti, Nick Nguyen and participants of the seminar organized by Massey University Albany campus on 8 May 2015 and Southwestern University of Finance and Economics on 1 July 2016; and comments received from the participants from the IFABS 2016 Barcelona Conference, “Risk in Financial Markets and Institutions: New challenges, New solutions” Universitat Autònoma de Barcelona (UAB), Casa Convalescència, Barcelona, June 1–3, 2016 and Asian Finance Association 2016 Annual Meeting, Bangkok, June 26–28, 2016. All errors are our own. ⁎ Corresponding author. E-mail addresses: [email protected] (Q. Liu), [email protected] (J. Luo), [email protected] (G.G. Tian).

http://dx.doi.org/10.1016/j.jcorpfin.2016.09.003 0929-1199/Crown Copyright © 2016 Published by Elsevier B.V. All rights reserved.

180

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

financing market, such as the corporate bond market and the banking system, is not well developed, with the consequence that firms get limited access to formal financing channels (Ge and Qiu, 2007). Numerous existing studies have provided reasons/arguments why firms receive or grant trade credit,1 and related empirical studies mainly link trade credit to other financial statement variables or characteristics of firms to find evidence to either support or reject those arguments.2 An increasing body of recent literature has begun to focus on how informal contracts and institutions, such as political connections and professional connections through business clubs, mitigate credit market frictions and facilitate firms' access to formal financial resources, such as bank loans and the equity market (Agrawal and Knoeber, 2001; Khwaja and Mian, 2005; Adhikari et al., 2006; Faccio, 2006; Faccio et al., 2006; Claessens et al., 2008; Le and Nguyen, 2009; Liu et al., 2013; Engelberg et al., 2012; Haselmann et al., 2013). It has been documented that asymmetric information and moral hazard problems are the major impediments to financial contracting because they create inefficiencies in the allocation of credit and raise the cost of providing credit (Haselmann et al., 2013). These frictions have a stronger influence on the supply of trade credit than on formal financing, such as bank loans, because suppliers take more risk in granting trade credit, as they neither charge any interest nor require any collateral or deposit (Wu et al., 2014). However, one unexplored question is whether and how the supply of trade credit is influenced by informal contracts of managers, such as their social and political networks through business clubs and political connections. In this paper, we attempt to shed light on this question by linking managerial personal professional and political connections to firms' access to informal financial resources – trade credit. In order to do so, we identify an important type of managerial professional connection that has never been studied before, i.e., managers who have working experience, or currently work, as an executive member of an industry association. We expect that by being an executive member of an industry association, such professional connections help firms' top executives to establish a good professional reputation and thus signal that any corporate decisions and repayment commitments made by those connected executives are reliable and trustworthy, and this kind of reputation and trust helps firms to reduce the credit risk faced by firms which offer credit. Professional connections seem to be more relevant to firms' access to trade credit for the following reasons: First, unlike formal financing, informal financing is always reputation and relationship-based (Ayyagari et al., 2010), which means that it can be greatly influenced by firms' social capital, such as managerial social connections. In addition, like business clubs in western countries,3 industry associations in China play an important social role, such as communication and collaboration between companies within the industry and with other related industries, which means that they enable the members, especially executive members, given that the duties of industry associations are conducted by the executive members, to build close social connections and managerial ties with other members. Such social connections and managerial ties have been identified as an important form of social capital for firms (Peng and Luo, 2000; Park and Luo, 2001; Luk et al., 2008; Haselmann et al., 2013) as they help executive members of associations to establish a better reputation and interpersonal trust with managers within the industry and related industries, such as upstream and downstream industries, which is vital to the supply of trade credit (Guiso et al., 2004; Wu et al., 2014). Furthermore, Wu et al. (2014) document that better regional social trust encourages firms to use trade credit. Thus we expect that interpersonal trust within an industry also plays an important role in helping firms to obtain external financial resources, and the role is more important for accessing trade credit than other forms of financial resources, given that trade credit is usually obtained from suppliers, who are usually from the same industry, or from upstream industries. Political connections, on the other hand, have been widely documented to bring various benefits to firms in accessing formal financial resources, such as bank loans and equity issuing, through rent seeking from government regulations or government lobbying (Berkman et al., 2010; Chen et al., 2011; Liu et al., 2013). However, they may be less important for firms' access to trade credit because government's intervention into financing resources mainly focuses on the formal financing market rather than the trade credit market, which is to say, government is less likely to have a direct influence on individual firms' decisions on whether to provide trade credit and how much to provide. Therefore, in this paper we hypothesize that managerial professional connections through industry associations, rather than political connections, help firms to access trade credit, and we provide comprehensive evidence to support our hypothesis by investigating the effect of managerial professional and political connections on the trade credit of Chinese listed firms. China is an ideal setting for our study for the following reasons: First of all, trade credit is more important in emerging markets such as China, where the formal financing markets, such as bank lending and the bonds market, are both poorly developed. And thus informal financing, such as trade credit, plays an even more important role in China than in the developed markets. Nevertheless, the Chinese economy is also characterized by a weak legal system that provides weak protection for creditors4 1 They are based on information asymmetries (Smith, 1987; Biais and Gollier, 1997), discrimination arguments (Brennan et al., 1988), monitoring advantages (Jain, 2001; Mateut et al., 2006), insurance (Cunat, 2007), product quality (Lee and Stove, 1993; Long et al., 1993), bankruptcy (Frank and Maksimovic, 2005; Wilner, 2000), opportunistic behaviour (Burkart and Ellingsen, 2004), externalities (Daripa and Nilsen, 2005) and trade-off between inventories and trade credit (Bougheas et al., 2009). 2 See, for example, Mian and Smith (1992), Rajan and Zingales (1995), Petersen and Rajan (1997), Ng et al. (1999), Demirguc-Kunt and Maksimovic (2002), Alphonse et al. (2003), Fisman and Love (2003), Giannetti (2003), Preve (2003), Burkart et al. (2005), Cunningham (2004), Atanasova and Wilson (2004), Love et al. (2005), Guariglia and Mateut (2006), and Cull et al. (2009). 3 Haselmann et al. (2013) document that such business clubs help firms to obtain bank loans, as banks lend primarily to firms within the same business club. 4 Although the legal system within China does not differ much from region to region, its implementation does. For example, according to the Supreme Court's judicial interpretation, the right of jurisdiction over all listed companies as defendants in a civil action belongs to the regional People's Intermediate Court. In this situation, the implementation of the legal system depends on whether the region has a well-developed free market with less government intervention in the economy. In regions that do not have a well-developed free market economy, listed companies are more likely to influence the local judicial department (the regional People's Intermediate Court) through networks or relationships, which dominate the Chinese economy. For example, Shanghai is considered to better implement the legal system than Tibet, due to the former's freer market and better developed economy.

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

181

(Peng, 2001; Kaoto and Long, 2005; Liu and Tian, 2012). Therefore suppliers of trade credit always face a high credit/default risk, and thus it is important to explore what is the channel through which Chinese firms seek to reduce the credit risk and obtain trade credit. Secondly, unlike developed markets, where the economy is dominated by formal contracts, in China, as a typical emerging market, relationship and personal connections (‘Guanxi’) play a pivotal role in the shaping and advancement of the daily business operations of firms, the so-called ‘Guanxi culture’ (Gu et al., 2008). Under this institutional setting, social ties with mangers of suppliers, buyers, competitors and other business intermediaries have a very strong influential power on corporate behaviours (Hsiung, 2013). Therefore, it is important to examine whether such managerial professional connections, as an important type of managerial social tie, have an influence on firms' ability to obtain trade credit, and the institutional environment of the Chinese market provides us with a unique setting to investigate this question. We expect that managerial professional connections should play an even more important role in helping Chinese firms to access trade credit because they enable firms to establish trust between suppliers and customers and thus mitigate the information asymmetry between suppliers and demanders of trade credit, and substitute for a good legal environment (Karlan, 2005). Thirdly, another important characteristic of Chinese firms is the co-existence of state controlled firms (SOEs) and non-stateowned firms (non-SOEs). The literature has documented that Chinese SOEs usually have preferential access to bank credit and the equity market, while non-SOEs are always discriminated against and always have limited access to formal financial resources (Li et al., 2008; Liu et al., 2013). This institutional setting in China allows us to explore whether SOEs and non-SOEs also have different access to trade credit and, more importantly, whether professional connections become a more important and effective mechanism for non-SOEs to overcome financial constraint than for SOEs. Therefore, the co-existence of SOEs and non-SOEs in China provides a unique setting for studying the competing effects of professional and political connections in accessing informal finance. Our empirical results show that managerial professional connections do play an important role in facilitating firms' access to trade credit, especially in non-SOEs, which have limited access to formal financing resources. It is possible that professional connections and political connections may co-exist, so we conduct a few tests to differentiate the effect of professional and political connections on trade credit, and our findings confirm that the positive relationship between professional connections and trade credit is mainly driven by managerial professional connections rather than the compounding effect of professional and political connections. In order to provide direct evidence to support our argument that professional connections help firms to access trade credit by establishing reputation and trust, we further investigate whether the relationship between professional connections and trade credit varies between firms with different levels of social trust and product market competition, and we document that the positive relationship between managerial professional connections and trade credit is strengthened by the high credit risk faced by suppliers, which is caused by a low level of social trust and high product market competition. By using alternative variables as proxies fort trade credit and professional connections, we find that managerial professional connections, as well as directors' professional connections, also result in more trade credit for financing purposes and more abnormal trade credit – more trade credit that is not related to the nature of goods. Moreover, we confirm that, compared to a lower level of professional connections and past professional connections, a higher level of connections and current connections plays a more important role in helping firms to obtain trade credit. It is worth noting that we also include managerial political connections in all our regressions. Our results show that political connections do have a statistically positive impact on firms' access to bank credit, which supports previous studies. However, the coefficients of political connections in our trade credit regressions are positive but all statistically and economically insignificant. This result suggests that managerial political connections play a relatively less important role in helping firms obtain trade credit. By using different approaches to address the potential endogeneity issue caused by omitted variable and reverse causality problems, we find that: (1) Managerial professional connections play a more important role in helping firms to obtain trade credit during the global financial crisis period, when the overall credit risk is higher and the information asymmetry between supplier and receiver of trade credit is more severe; (2) We document that firms which have more employees and whose top executives have a higher educational background are more likely to be professionally connected, and, more importantly, we confirm the positive relationship between professional connections and trade credit by using the Heckman two-stage regression model; (3) It is possible that the difference in trade credit in firms with and without professional connections is caused by other factors. We address this issue using the propensity score matching method, and we find that the difference in trade credit is statistically significant when we match our sample based on different criteria; (4) We compare the level of trade credit before and after firms establish or lose professional connections, and we find that firms' trade credit increases significantly after the professional connections are established, and decreases, though insignificantly, after firms lose the connections. Overall, the above endogeneity tests suggest that our results are robust after controlling for the potential endogeneity issue. Lastly, by conducting a few robustness tests, we document that (1) firms with professional connections also offer more trade credit to other firms. However, the net position of trade credit that firms receive is still significantly higher in firms with professional connections after deducting the trade credit that firms offer to their customers. (2) Managerial professional connections have only a marginal effect on firms' access to formal financial resources – bank credit; and (3) our results are not driven by the large firm size effect. Our study contributes to the literature in the following ways. First of all, we contribute to the literature on the implications of managerial characteristics for corporate financial decisions. Previous studies have provided substantial evidence that managerial characteristics, such as early life experience, managerial expertise, overconfidence, have an important influence on corporate policy (Guner, Malmendier and Tate, 2008; Malmendier and Tate, 2008; Malmendier et al., 2011) and that managers' political

182

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

connections play an important role in helping firms obtain external formal financing resources, such as bank loans and equity issuing (Agrawal and Knoeber, 2001; Khwaja and Mian, 2005; Liu et al., 2013), but there is a lack of evidence about whether managerial personal connections, such as their professional connections, have any influence on firms' access to informal financing resources. For the first time in the literature, this study documents that managerial personal professional connections have an important influence on firms' access to trade credit. Thus we complement the literature in this area. In addition, by showing that managers' political connections do not have an influence on trade credit, this study differentiates the role played by different types of managerial characteristics. In particular, while previous studies provide substantial evidence that managers' political connections are important in helping firms to obtain formal financial resources, our findings indicate that managerial professional connections are more relevant for firms in obtaining informal financing than are managerial political connections. By doing so, we provide evidence that different types of managerial social connections play different roles when firms seek to access different financial resources. Our study also contributes to the literature on trade credit because previous studies on trade credit mainly focus on how trade credit is influenced by financial statement variables or firm/product characteristics (Lee and Stove, 1993; Petersen and Rajan, 1997; Nilsen, 2002; Mateut et al., 2006; Cunat, 2007; Giannetti et al., 2011), while very few studies investigate how trade credit is influenced by managers' personal characteristics. Our paper fills this gap by linking managerial professional connections to firms' access to trade credit. Finally, we make a contribution to the literature on social trust. Recent studies suggest that the local social trust environment has an important influence on firms' corporate policy. Jha and Chen (2015) find that firms which locate in a region with a high level of mutual trust pay fewer audit fees, using a sample of US firms, and Wu et al. (2014) document that Chinese firms that locate in regions with a high level of social trust and a reliable legal environment are more likely to use trade credit. Our paper complements their papers by providing evidence that managerial interpersonal trust through professional connections, rather than regional social trust, also helps firms to obtain and trade credit. The remainder of the paper is organized as follows. Section 2 develops our main hypotheses. Section 3 describes the sample selection, construction of variables and summary statistics. Section 4 presents our main empirical results and robustness tests. Section 5 concludes the paper. 2. Development of hypotheses 2.1. Professional connections through industry associations and firms' access to trade credit Based on our analysis in Section 1, our main hypothesis is that managerial professional connections through industry associations should have a positive influence on the trade credit that firms can obtain, because their professional working experience enables managers to establish a better reputation and interpersonal trust within the associated industries, thus alleviating the information asymmetry and reducing the credit risk faced by suppliers. This may increase the probability that suppliers provide trade credit to firms with professional connections, while political connections should play a less important role in helping firms to access trade credit, although they enable firms to better access bank credit (Khwaja and Mian, 2005). Therefore, we develop our base hypothesis as follows: H1. Managerial professional connections through industry associations are significantly positively related to firms' access to trade credit, while managerial political connections have a less significant influence on firms' access to trade credit. 2.2. The effect of managerial professional connections on firms' access to trade credit in SOEs and non-SOEs If, as expected, professional connections help firms to access trade credit by establishing reputation and social trust, one rational extension of this conjecture is that professional connections should play a more important role when firms have a greater need for trade credit and when the information asymmetry between credit suppliers and credit recipients is more severe. We thus expect a stronger positive relationship between professional connections and firms' access to trade credit in non-SOEs and a weaker relationship in SOEs. China is characterized by a state-dominated banking system, which is always blamed for misallocating credit by lending primarily to state-controlled firms (Cull and Xu, 2003; Liu et al., 2016). Under this institutional setting, Chinese non-SOEs have to rely more on trade credit to overcome financial constraint. However, compared to SOEs, due to the lack of implicit government guarantee, it is usually harder for non-SOEs to obtain trade credit, which means non-SOEs have to rely on managerial professional connections to establish social trust and obtain trade credit. In other words, managerial professional connections should play a more significant role in non-SOEs than in SOEs. We thus further hypothesize that: H2. The positive relationship between managers' professional connections and firms' access to trade credit is more pronounced in non-SOEs and less pronounced in SOEs. 2.3. How do managerial professional connections help firms to gain trade credit? We expect that the major channel through which professional connections help firms to obtain trade credit is reputation, which is to say, by being an executive member of an industry association, professional connections signal that the top executive

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

183

of a firm is reputable and thus corporate decisions made by that top executive are more trustworthy. And thus, professional connections may mitigate the inverse effect of information asymmetry and reduce the credit risk faced by the suppliers of trade credit. Therefore, it is expected that managerial professional connections should play a more important role when the credit risk is higher. It has been documented that trade credit is a ‘free’ resource and the credit-receiving firms are not required to pay interest, but at the same time, the suppliers take a high credit risk for granting trade credit because trade credit is an implicit and informal financing contract with no collateral standing behind the transaction and no guarantees from third parties or financial intermediaries (Wu et al., 2014). Thus, whether a firm could be granted trade credit largely depends on the credit risk between suppliers and customers, or in other words, the extent to which the suppliers trust the customers (Guiso et al., 2004). Following Wu et al. (2014) we argue that good regional social trust reduces the credit risk faced by credit suppliers and thus facilitates firms' access to trade credit. We therefore expect that managerial professional connections should play a more important role in facilitating firms' access to trade credit in firms that locate in regions with lower social trust (and higher credit risk), which is to say, managerial professional connections substitute good social trust, and develop the following hypothesis: H3a. The positive relationship between managers' professional connections and firms' access to trade credit is more pronounced in credit-receiving firms which operate in regions with a lower level of social trust. In addition, the level of credit risk is also greatly influenced by the product market competition in which firms operate. In particular, in industries with greater competition, suppliers may face more credit risk when offering trade credit and are thus less likely to grant trade credit to their customers. This variation in credit risk could also influence the effect that professional connections have on trade credit. When the credit risk is low, firms can get access to trade credit easily; when the credit risk is high, firms may not be able to obtain trade credit easily and managerial professional connections through industry associations become more important. Therefore, we develop the following hypothesis: H3b. The positive relationship between managerial professional connections and firms' access to trade credit is more pronounced in credit-receiving firms which operate in more competitive industries.

3. Research design 3.1. Data collection and sample selection In order to conduct our study, we collected all Chinese listed firms in both the Shanghai stock exchange and the main board of the Shenzhen stock exchange.5 The data used in this study is mainly collected from a series of datasets provided by the Chinese Stock and Market Accounting Research (CSMAR) database from 2006 to 2012. Firms' financial information is collected from the Chinese Listed Firm Annual Report Database. Information about the managers' backgrounds and working experience is collected partly from the Chinese Listed Firm Corporate Governance Database, which discloses the biographical details of all executives, including the chairman, CEO and all board members and other senior executives. We combine this information with other manually collected information from the public media. Information about ownership structure and other financial variables is collected from the Chinese Listed Firm Shareholder Analysis Database and Chinese Listed Firm Financial Variables Analysis Database. The following abnormal observations are excluded: (1) financial firms (firms with unique accounting standards and special financial characteristics); (2) ST (special treatment) firms or negative-equity firms (financially distressed firms); (3) firms with a sales growth rate of more than 1 or less than 0; (4) cross-board listing firms, which means firms that are listed both in the Shanghai/Shenzhen stock exchanges and overseas markets, or firms that issue both A-shares and B-shares; (5) delisted firms or newly listed firms; and (6) firms whose relevant data are not complete or cannot be acquired. To minimize the effect of outliers, we trimmed our sample at 1% on each variable in each tail. The final sample consists of 655 firms with 3986 firm-year observations from 2006 to 2012. 3.2. Measuring variables 3.2.1. Professional connections through industry associations and political connections To identify whether the firms' top executives/board members have professional connections, we first collect the information about the managers' backgrounds and working experience from the relevant databases, to identify whether the top executives and board members currently serve, or have working experience, as secretary, associate secretary, chairman, associate chairman, director or executive director, of an industry association. Of all the industry associations in China, we: (1) select those associations which are registered in the Chinese Social Organization website, a department of the Ministry of Civil Affairs6; and (2) search the name of the relevant association using Baidu search engine, and visit the website to select the ‘active’ associations in our sample

5 Firms listed in the Small and Medium Sized Enterprise board (SME) and Growth Enterprise Market board (GEM) within the Shenzhen Stock Exchange are not included because they are mostly young and expanding companies which are not comparable to firms listed in the main board. 6 The website of the Chinese Social Organization: http://www.chinanpo.gov.cn/index.html.

184

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

period. Finally we obtain a total number of 145 active industry associations in China. Then, we define the top executives' professional connections (PROCON) as a dummy equal to 1 if the CEO/chairman of a firm is an executive member of an industry association in a particular year. Please note that this is our key regression variable; for robustness of our results, we may also create some other variables to capture different characteristics of professional connections, the definitions of which will be provided in the following sections when necessary. In the spirit of the definitions of political connections used by Fan et al. (2007), Wu et al. (2010) and Liu et al. (2013), a firm is defined as having political connections (PC dummy equals 1) if either its chairman or general manager/CEO currently serves or

Table 1 Sample distribution and summary statistics. Panel A Distribution of managerial professional connections and firms' access to trade credit based on year. Obs. represents the number of observations in each year; Obs. with PROCON represents the number of observations with professional connections in each year; Percentage (%) is the percentage of firms with professional connections in each year. Ave. CREDIT (%) is the average trade credit that firms obtain (accounts payable/total assets) in each year.

2006 2007 2008 2009 2010 2011 2012 Total

Obs.

Obs. with PROCON

Percentage (%)

Ave. CREDIT (%)

487 508 544 569 605 627 646 3986

85 103 168 175 173 184 107 995

17.45 20.28 30.88 30.76 28.60 29.35 16.56 24.96

9.50 8.15 7.16 10.32 20.18 9.94 7.95 11.02

Panel B Distribution of professional connections through industry associations and firms' access to trade credit based on year Obs. represents the number of observations in each industry; Obs. with NETWORK represents the Obs. number of observations with social networks through industry association in each industry; Percentage (%) is the percentage of firms with social connections with industry association in each industry. Ave. CREDIT (%) is the average trade credit that firms obtain (accounts payable/total assets) in each industry.

Obs. with PROCON NETWORK

Percentage Ave. (%) CREDIT (%)

Agriculture Mining Manufacture Electricity, gas and water Construction Transportation Information technology Wholesale and retail Real estate Social service Broadcasting and culture Conglomerate Total

66 73 303 66 72 46 59 112 99 46 20 33 995

41.67 36.67 29.49 9.71 18.03 8.43 13.85 30.36 42.86 53.85 27.27 27.78 24.96

146 183 949 627 371 505 395 341 213 79 67 110 3986

10.85 10.26 23.24 8.92 9.99 10.07 10.38 10.57 8.76 10.73 9.04 10.30 11.02

Panel C Summary statistics of regression variables Variable

N

MEAN

STDEV

MIN

MEDIAN

MAX

CREDIT S_CREDIT PROCON PC NON-SOE TRUST MARKETIZATION HHI BANKLOAN (%) SIZE (Million Yuan) AGE GROWTH (%) TOBIN'Q LEVERAGE (%) OPCASH (%) FIXASSET ROA EBIT CEOEDU CEOCOMP CEOEXP

3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3986 3037 3037 3037

0.11 0.10 0.25 0.37 0.42 2.541 7.01 0.11 20.03 22,190 12.97 0.54 1.62 0.54 19.39 0.26 0.04 0.58 3.43 12.68 1.52

0.08 0.08 0.19 0.45 0.49 3.811 1.86 0.10 2.05 11,800 3.29 0.16 1.06 0.16 1.520 0.19 0.05 2.2 0.84 1.03 1.85

0 0 0 0 0 0.1 3.15 0.03 9.11 18,270 1 0.01 0.50 0.11 12.29 0 −1.15 −3.7 1 0 0.33

0.08 0.08 0 1 1 0.7 6.88 0.08 20.23 22,000 13 0.54 1.26 0.54 19.74 0.24 0.03 0.23 4 12.77 1

0.60 0.45 1 1 1 22.7 11.8 1 25.29 277,500 29 0.99 16.01 0.85 25.79 0.95 0.52 11 6 17.68 19

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

185

formerly served in the government or military, or serves/served as a deputy of the National/Provincial People's Congress or the People's Political Consultative Conference.

3.2.2. Trade credit and bank credit Following previous studies, such as Petersen and Rajan (1997), Fisman and Love (2003), Giannetti et al. (2011) and Wu et al. (2014), this study uses accounts payable and receivable to measure the total trade credit a firm receives or grants; both are scaled by total assets. This study mainly focuses on the trade credit that firms receive, so our key variable, firms' access to trade credit (CREDIT), is measured by total accounts payable/total assets. We also measure firms' supply of trade credit (S_CREDIT) as total accounts receivable/total assets. And the net position of trade credit (NET_CREDIT) is measured by the trade credit a firm receives minus the trade credit a firm supplies. Following previous studies, such as Firth et al. (2009), we measure bank credit (BANK) as the ratio of bank loans to total assets.

3.2.3. Control variables Our firm-specific control variables are defined as in previous studies, such as Ge and Qiu (2007), Love et al. (2005), and Wu et al. (2014), to capture the effect from other factors on trade credit. Detailed definitions of all variables used in this study are reported in Appendix A.

3.3. Descriptive analysis and univariate test 3.3.1. Descriptive analysis Table 1 presents the sample distribution and descriptive statistics for our sample firms. Panels A and B report the distribution of professional connections/trade credit based on year and industry respectively. The results in panel A show that on average 24.96% of our sample firms have professional connections through industry associations. The yearly distribution shows that the percentage of firms that have professional connections increases sharply in 2008 and stays at a high level during the global financial crisis period (2008 to 2009). The general distribution of trade credit is quite similar to that of professional connections, suggesting that a positive relationship exists between the two variables. As shown in panel B, the distribution of professional connections also varies between industries: in general, firms in the agriculture, mining, manufacture, wholesale and retail, real estate, and social service industries are more likely to have professional connections through industry associations, and they also use relatively more trade credit. Monopolized industries, such as electricity, gas and water service and transportation, have the lowest level of professional connections and use less trade credit. The probable reason is that the market competition in those industries is low, so firms in those industries do not rely on trade credit and do not need professional connections to help their trade credit financing. Panel C of Table 1 reports the summary statistics of our main regression variables. On average, Chinese listed firms receive 11% trade credit (accounts payable) and offer 10% trade credit (accounts receivable) to other firms.7 Compared with the average bank loan ratio of 20.03%, our results indicate that trade credit is another important channel through which firms obtain financial resources. The level of obtaining and offering trade credit (accounts payable and accounts receivable) is similar to previous studies, such as Wu et al. (2014) and Lin and Chou (2014). Other main results show that 25% of our sample firms have professional connections through industry associations and 37% of them have political connections, suggesting that professional and political connections are quite prevalent in Chinese listed firms.

3.3.2. Univariate test Table 2 presents the univariate test results, where we compare the average trade credit between firms with and without professional as well as political connections. Our results in panel A show that firms with professional connections use12.01% of trade credit, while firms without such connections use only 10.05% of trade credit, and the difference (1.96%) is statistically significant at the 5% level of significance. This result confirms our hypothesis H1 that managerial professional connections help firms' access to trade credit. When dividing our sample of firms into subsamples of SOEs and non-SOEs, we further show that managerial professional connections play a more important role in non-SOEs than in SOEs, which is consistent with our hypothesis H2. From panel B, we find that managerial political connections play an insignificant role in helping firms' access to trade credit. In order to further provide evidence for our hypotheses H3a and H3b, we repeat our univariate test based on different subsamples of firms. As expected, we find that managerial professional connections bring firms significantly more trade credit in firms that locate in regions with a low level of social trust and firms operating in industries with high market competition, which is consistent with our hypotheses H3a and H3b.

7 The reason for the difference in percentage between average trade credit received and offered is because listed firms may receive or offer credit to non-listed firms or individuals.

186

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 2 Univariate tests. Panel A Trade credit between firms with and without professional connections. This panel presents firms' access to trade credit between firms with and without professional connections. NON-PROCON/PROCON refer to firms without/with professional connections. *, ** and *** represent significance at the 10%, 5% and 1% levels. NON-PROCON

FULL SAMPLE SOE NON-SOE

PROCON

DIFFERENCE

MEAN (%)

MEDIAN (%)

MEAN (%)

MEDIAN (%)

T-value (%)

Z-value (%)

10.05 9.94 10.45

9.37 8.67 10.18

12.01 11.24 12.74

11.05 11.14 11.56

1.96** 1.30 2.29**

1.68*** 2.47* 1.38***

Panel B Trade credit between firms with and without political connections. This panel presents firms' access to trade credit between firms with and without political connections. NPC/PC refer to firms without/with political connections. *, ** and *** represent significance at the 10%, 5% and 1% levels. NPC

FULL SAMPLE SOE NON-SOE

PC

DIFFERENCE

MEAN (%)

MEDIAN (%)

MEAN (%)

MEDIAN (%)

T-value (%)

Z-value (%)

10.43 9.92 11.40

9.51 9.78 10.72

12.04 11.43 12.57

11.16 11.02 11.92

1.61 1.51 1.17

1.65 1.24 1.20

Panel C Trade credit between firms with and without social connections and different classification variables. NON-PROCON/PROCON refer to firms without/with professional connections. *, ** and *** represent significance at the 10%, 5% and 1% levels. NON-PROCON

LOW TRUST HIGH TRUST LOW COMPETITION HIGH COMPETITION

PROCON

DIFFERENCE

MEAN (%)

MEDIAN (%)

MEAN (%)

MEDIAN (%)

T-value (%)

Z-value (%)

8.36 11.15 9.41 7.74

8.23 10.78 6.34 5.77

10.42 12.38 9.11 12.40

10.31 11.84 8.39 8.35

2.06** 1.23 −0.30 4.66***

2.08*** 1.06 2.05 2.58***

Table 3 The effect of managerial professional and political connections on firms' access to trade credit. This Table reports the effect of managers' professional and political connections on firms' ability to obtain trade credit. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; PC is a dummy for managers' political connections. Columns 1, 2 and 3 report the results based on the full sample, SOE subsample and non-SOE subsample separately. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT ALL

SOE

NON-SOE

PROCON

0.014** (2.260) 0.002 (1.275) 0.026* (12.71) 0.004 (1.530) −0.162*** (−16.301) 0.003 (1.643) 0.011*** (6.842) 0.161*** (5.077) 0.322*** (30.158) −0.170 (−1.537) Controlled in all regressions Controlled in all regressions

0.013 (0.926) 0.001 (1.451) 0.021*** (11.414) 0.003 (0.716) −0.140*** (−10.364) 0.002 (1.072) 0.023*** (4.366) 0.133*** (5.026) 0.469*** (10.594) −0.189 (−1.125)

0.016** (2.292) 0.037 (1.504) 0.053*** (11.586) 0.005* (1.723) −0.215*** (−7.874) 0.011 (0.081) 0.002*** (3.866) 0.214 (1.430) 0.160*** (11.877) −0.2438 (−0.698)

χ(2) = 37.64 2570 0.502 125.33**

1416 0.549 118.23**

PC SIZE AGE FIXASSET EBIT OPCASH ROA LEV Constant YEAR INDUSTRY Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted – R2 F

3986 0.57 160.91***

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

187

Table 4 Addressing the co-existence of professional and political connections. Panel A Correlation coefficient of professional connections and political connections. This panel presents the correlation coefficient between managerial professional connections and political connections. PROCON is a dummy for managers' professional connections; PC is a dummy for managers' political connections. ** represents significance at the 5% level. PROCON 1 0.328**

PROCON PC

PC 1

Panel B Co-existence of professional and political connections by year. This panel reports the extent of co-existence of professional and political connections by year. Firms with PROCON and Firms with PC refer to the total number of firms with professional connections and political connections separately. Within firms with professional connections, we identify whether the connection is central level, provincial level and city level. Firms with double connections refers to firms with both professional and political connections. The last two rows report the percentage of firms with double connections relative to the total number of firms with professional/political connections.

Firms with PROCON - Central level -Provincial level -City level Firms with PC Firms with double connections Percentage of firms with double connections to firms with PROCON (%) Percentage of firms with double connections to firms with PC (%)

2006

2007

2008

2009

2010

2011

2012

Total

85 41 30 14 106 34 40 32.08

103 49 35 19 110 36 34.95 32.73

175 82 63 30 108 54 30.86 50

168 82 56 30 113 49 29.17 43.36

173 80 62 31 106 57 32.95 53.77

184 80 72 32 115 64 34.78 55.65

107 49 41 17 110 42 39.25 38.18

995 463 359 173 768 336 33.77 43.75

Panel C Univariate test of firms with different connections. This panel compares the trade credit between firms with different characteristics. Group 1 refers to firms without any connections. Group 2 refers to firms with professional connections but without political connections, while Group 3 is firms with both professional and political connections. CREDIT

Gropu1 Group2 Group3

Mean

0.101 0.119 0.122

Median

0.094 0.109 0.115

Difference (T-test for significance) Group2 vs Group1

Group3 vs Group2

Group2-Group1 vs Group3-Group2

0.018** (T = 2.247)

0.003 (T = 1.049)

0.015** (T = 2.138)

Panel D The effect of professional connections on trade credit in firms with and without political connections. This panel presents the effect of professional connections on trade credit of firms with and without political connections. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; PC is a dummy for managers' political connections. Columns 1, 2 and 3 report the results based on the full sample, firms without and firms with political connections separately. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

PROCON PC SIZE AGE FIXASSET EBIT OPCASH ROA LEV CONSTANT YEAR INDUSTRY Test for difference of the coefficient of PROCON in NPC and PC

CREDIT ALL

NPC

PC

0.014** (2.260) 0.002 (1.275) 0.026* (12.71) 0.004 (1.530) −0.162*** (−16.301) 0.003 (1.643) 0.011*** (6.842) 0.161*** (5.077) 0.322*** (30.158) −0.170 (−1.537) Controlled in all regressions Controlled in all regressions

0.017*** (2.580)

0.0039 (1.146)

0.029* (11.735) 0.006 (0.931) −0.413*** (−13.801) 0.015*** (9.547) 0.011* (1.713) 0.010*** (9.681) 0.328*** (14.865) 0.913** (2.119)

0.0169 (0.958) 0.0119 (0.182) −0.0922 (−0.805) 0.001* (1.851) 0.0015** (2.557) 0.310* (1.909) 0.125*** (8.753) 0.0179 (0.382)

χ(2) = 23.77 (continued on next page)

188

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 4 (continued) Panel D The effect of professional connections on trade credit in firms with and without political connections. This panel presents the effect of professional connections on trade credit of firms with and without political connections. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; PC is a dummy for managers' political connections. Columns 1, 2 and 3 report the results based on the full sample, firms without and firms with political connections separately. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT

Observations Adjusted-R2 F

ALL

NPC

PC

3986 0.57 160.91***

3218 0.424 119.9**

768 0.482 112.06**

3.4. Regression models In order to examine the effect that professional and political connections have on firms' access to trade credit, the following baseline regression model is established: CREDIT i;t ¼ ∝ þ β1  PROCONi;t þ β2  PC i;t þ β3  X i;t þ Year and Industry dummies

ð1Þ

In Eq. (1), the dependent variable (CREDIT) is firms' access to trade credit measured by total accounts payable/total assets. The main independent variables are managers' professional connections (PROCON) and political connections (PC). X is a vector of control variables; definitions are reported in Appendix A. Year and industry dummies are included in all regression models in order to control for firm-specific and industry-specific effects. Please note that Eq. (1) is only our baseline model; additional variables will be added in our baseline regression model to interact with our key independent variable (PROCON) to investigate whether the relationship between trade credit and managerial professional connections is weakened or strengthened by other factors. Table 5 The effect of professional connections and social trust on trade credit. This Table presents the effect of professional connections and social trust on trade credit. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; TRUST is the regional level of social trust following the survey conducted by the “Chinese Enterprise Survey System” in 2001. Columns 1, 2 and 3 report the results based on the full sample, SOE subsample and non-SOE subsample separately. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT ALL

SOE

NON-SOE

PROCON

0.020** (2.187) 0.039* (1.889) −0.017*** (−2.526) 0.002 (1.211) 0.016 (1.630) 0.006*** (3.341) −0.085*** (−13.825) 0.003 (1.181) 0.010 (0.715) 0.158*** (7.346) 0.116*** (17.803) −0.020 (−0.954) Controlled in all regressions Controlled in all regressions P = 0.032 3986 0.402 197.9**

0.018 (1.422) 0.006 (0.270) −0.013 (−0.116) 0.001 (0.897) 0.015 (0.132) 0.002** (2.354) −0.069*** (−8.747) 0.002 (1.330) 0.020 (1.018) 0.132*** (6.693) 0.104*** (12.674) 0.010 (0.385)

0.022** (2.021) 0.089** (2.250) −0.020*** (−2.406) 0.007 (1.327) 0.003* (1.646) 0.006** (2.076) −0.092*** (−7.649) 0.001 (1.208) 0.001 (1.484) 0.162*** (4.771) 0.140*** (12.312) −0.064 (−1.536)

P = 0.001 2570 0.355 43.68**

P = 0.026 1416 0.332 153.7***

TRUST PROCON*TRUST PC SIZE AGE FIXASSET EBIT OPCASH ROA LEV CONSTANT YEAR INDUSTRY P-value interaction Observations Adjusted – R2 F

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

189

Table 6 How the product market competition influences the effect of professional connections on firms' access to trade credit? The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; HHI is the Herfindahl-Hirschman Index, which measures the product market competition. Detailed definitions for all variables are reported in Appendix A. Columns 1, 2 and 3 report the results based on the full sample, SOE subsample and nonSOE subsample separately. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT ALL

SOE

NSOE

PROCON

0.014*** (4.328) −0.234** (−2.337) −0.588** (−2.385) 0.009* (1.751) 0.022 (1.508) 0.0365 (1.106) −0.102*** (−15.069) 0.014 (1.443) 0.055*** (5.634) 0.085*** (4.095) 0.101*** (14.598) −0.083 (−1.588) Controlled in all regressions Controlled in all regressions 3986 0.385 P = 0.001

0.013 (1.567) −0.181 (−1.452) −0.612 (−1.518) 0.007 (1.583) 0.028 (1.540) 0.0438 (1.082) −0.085*** (−10.455) 0.004 (1.241) 0.054*** (4.488) 0.133*** (4.984) 0.090*** (10.781) −0.056 (−1.057)

0.015* (1.809) −1.809*** (−5.889) −0.276* (−1.943) 0.010* (1.786) 0.042 (1.284) 0.0702 (1.085) −0.118*** (−8.240) 0.002 (0.041) 0.064*** (3.505) 0.070** (1.967) 0.183*** (13.185) −0.042 (−0.669)

2570 0.402 P = 0.0416 χ(2) = 31.45 24.96

1416 0.463 P = 0.000

HHI HHI*PROCON PC SIZE AGE FIXASSET EBIT OPCASH ROA LEV Constant YEAR INDUASTRY Observations Adjusted-R2 P-value for interaction Test for difference of the coefficient of PROCON in SOE and NON-SOE F

35.13

18.93

4. Empirical results and analysis 4.1. The effect of professional and political connections on firms' access to trade credit We first run our baseline regression model to examine whether and how managers' professional connections and political connections influence firms' access to trade credit. The results are reported in Table 3. Our results in column 1 of Table 3 show that in general, managerial professional connections through industry associations have a positive effect on firms' access to trade credit, based on the results that the estimated coefficient of PROCON is 0.014 and is statistically significant at the 5% level of significance. This result indicates that managers' professional connections help firms to obtain trade credit. In particular, we find that a 1 % increase of professional connections results in about a 0.014% increase of trade credit ratio (equivalent to about a 3.1 million RMB increase of trade credit). Interestingly, the estimated coefficient of managers' political connections (PC) is also positive but is statistically insignificant, suggesting that, although managers' political connections help firms to obtain formal financial resources, such as bank loans or access to the equity market, which was found in other studies (Li et al., 2008; Liu et al., 2013), they do not have a direct influence on firms' access to informal financial resources, such as trade credit. The probable reason is: politically connected managers facilitate firms' access to formal financial resources mainly through their ability to seek rent from government-owned banks/government regulations (Khwaja and Mian, 2005; Li et al., 2008; Liu et al., 2013), but compared to those formal financial resources, trade credit is not directly associated with government regulations. The results confirm our hypothesis H1. In terms of firm-specific control variables, we find that firm age is statistically significantly positively related to firms' access to trade credit, while firm size is statistically significantly negative; the results are similar to Ge and Qiu (2007), who argue that the influence that firm age and size have on firms' access to trade credit may depend on two forces: on the one hand, older and larger firms may have better established records and reputation and may find it easier to finance through trade credit. On the other hand, older and larger firms might have better banking relationships and thus have less need for trade credit. Our results suggest that larger firm size and older firm age both help firms to obtain trade credit but older firm age helps firms to obtain trade credit. The ratio of fixed assets in total assets is statistically significantly negatively associated with firms' access to trade credit, a result

190

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 7 The effect of professional connections on trade credit: alternative measures of key variables. This Table presents the effect that managerial professional connections/directors' professional connections have on firms' access to the trade credit/financing component of trade credit. The dependent variable CREDIT is the total accounts payable to total assets; FIN_CREDIT is the financing component of trade credit, following Ge and Qiu (2007). PROCON is a dummy for managers' professional connections; DIR_PROCON is a measurement for directors' professional connections. Detailed definitions for all variables are reported in Appendix A. Columns 1 (4), 2 (5) and 3 (6) report the results based on the full sample, SOE subsample and non-SOE subsample separately. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT SOE

FIN_CREDIT NSOE

PROCON DIR_PROCON OPCAH FIXASSET AGE SIZE LEV ROA PC EBIT CONSTANT YEAR INDUSTRY Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted-R2 F

0.045 0.054** (0.823) (2.036) 0.017 0.019 (1.337) (1.417) −0.012 −0.002 (−0.355) (−1.113) 0.080** 0.014 (2.005) (0.330) 0.049*** 0.015 (2.688) (0.561) 0.115*** 0.141*** (13.691) (13.716) 0.089*** 0.137*** (3.126) (4.942) 0.007 0.002 (1.530) (0.582) 0.010 0.003 (0.029) (0.198) −0.154*** 0.049 (−5.384) (1.001) Controlled in all regressions Controlled in all regressions χ(2) = 40.95 2570 1416 0.350 0.400 28.3** 41.43**

SOE

NSOE

0.018 (0.677)

0.014*** (2.364)

0.034* (1.825) −0.004 (−0.344) 0.011* (1.878) 0.038 (1.378) 0.133*** (10.569) 0.054 (1.263) 0.017 (0.878) 0.004 (0.866) −0.012 (−0.286)

0.010*** (4.595) −0.001 (−1.191) 0.038*** (5.469) 0.042 (0.988) 0.171*** (10.246) 0.072 (1.609) 0.022 (0.302) 0.002** (2.072) −0.132* (−1.647)

χ(2) = 37.28 2570 1416 0.275 0.399 29.59* 94.99**

SOE

NSOE

0.002 (1.539) 0.034* (1.781) −0.002 (−0.332) 0.011* (1.872) 0.037 (1.350) 0.132*** (10.480) 0.052 (1.214) 0.017 (0.883) 0.002 (0.852) −0.012 (−0.283)

0.045*** (2.335) 0.011*** (4.774) −0.003 (−1.114) 0.040*** (5.626) 0.078* (1.837) 0.173*** (10.341) 0.077* (1.714) 0.024* (0.661) 0.021* (1.669) −0.067 (−0.841)

χ(2) = 39.71 2570 1416 0.276 0.395 29.67*** 93.17**

that is consistent with Wu et al. (2014). Firms with more fixed assets have better access to formal financing because those assets can be used as collateral when applying for bank loans, and thus those firms have less need for trade credit. As in Ge and Qiu (2007) and Wu et al. (2014), operating cash flow and return on assets are both significantly positively related to firms' access to trade credit, a reasonable result because those firms usually have a stronger ability to repay their payables, and so have a lower default risk. Leverage is strongly positive, partly because accounts payable are themselves part of leverage, so firms with a high level of accounts payable also have high leverage. In addition, when we conduct our regression in SOEs and non-SOEs, we find from columns 2 and 3 that managers' professional connections have a greater and more significant influence on trade credit in non-SOEs than in SOEs (the regression coefficient of PROCON is 0.016 for non-SOEs and is significant at 5% significance, while it is 0.013 for SOEs and statistically insignificant). Thus our hypothesis H2 is also supported.

4.2. Addressing the issue of the co-existence of professional and political connections Our above results suggest that managers' professional connections and political connections have different functions in helping firms to access trade credit. By saying so, we assume managers' professional and political connections are independent and do not have any overlapping effects. However, in reality, it is possible that one person could have both professional and political connections; in other words, the two types of connections may be co-existent. Thus we attempt to address this issue in this section. We first calculate the correlation coefficient of the two variables to see the extent of co-existence of the two types of connections. The results are reported in panel A of Table 4. The correlation coefficient between the two types of connections is positive (0.328) and is statistically significant at the 5% level, confirming the co-existence of the two types of connections. The co-existence of the two types of connections is also confirmed by our results in panel B of Table 4, which show on average 33.77% of firms with professional connections (or 43.75% of firms with political connections) have double connections. Then, the next related question is: given the co-existence of the two connections, whether the positive relationship between managerial professional connections and trade credit is caused by only the professional connections or a compounding effect of both connections. To answer this question, we divide our full sample of firms into three groups: firms without any connections (Group 1), firms with professional connections but without political connections (Group 2) and firms with both professional

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

191

Table 8 The effect of professional connections on abnormal trade credit. This Table presents the effect that managerial professional connections have on abnormal trade credit. The dependent variable ABNORMAL_CREDIT is the measure for abnormal trade credit (see in footnote 7 for the calculation of abnormal trade credit). PROCON is a dummy for managers' professional connections. Detailed definitions for all variables are reported in Appendix A. Columns 1, 2, 3 and 4 report the results based on the full sample, SOE subsample and non-SOE subsample separately. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

ABNORMAL_CREDIT ALL

PROCON PC SIZE ROA OPCASH FIXASSET EBIT LEV AGE CONSTANT YEAR INDUSTRY Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted R-squared F

ALL

0.971* (1.902)

0.989* (1.827) 1.008 (1.244) 3.006*** 2.995*** (10.722) (10.689) 3.18** 3.09** (2.137) (2.193) 0.476** 0.455** (2.457) (2.351) −0.439 −0.547 (−0.351) (−0.438) 1.96 1.87 (1.006) (1.068) 5.20*** 5.17*** (11.081) (11.072) 0.0930 0.0995 (1.433) (1.532) −7.62*** −7.89*** (−9.539) (−9.288) Controlled in all regressions Controlled in all regressions 3986 0.342 20.53**

3986 0.343 20.34***

SOE

NSOE

0.807 (0.117) 0.960 (0.715) 3.037*** (10.918) 3.721* (1.811) 0.233* (1.957) −0.641 (−0.490) 0.922 (0.850) 6.193*** (10.598) 0.0454 (1.162) −7.92*** (−9.291)

1.072** (2.387) 1.334 (1.410) 2.002*** (9.740) 2.69*** (3.280) 0.486*** (6.209) −0.244 (−0.337) 1.92 (0.311) 2.76*** (11.647) 0.163 (0.169) −7.73*** (−9.325)

χ(2) = 36.15 2570 0.343 20.07**

1416 0.395 20.96**

and political connections, i.e., firms whose top executives actually have two types of connections (Group 3), and we compare the level of trade credit between the three groups of samples. The results are reported in panel C of Table 4. We find that, compared to firms without any connections (Group 1), firms with only professional connections (Group 2) use a significantly higher level of trade credit, which is consistent with our main argument. Although firms with double connections (Group 3) are found to have slightly higher trade credit compared to firms with only professional connections (Group 2), the difference is marginal and statistically insignificant, suggesting that our results are not driven by political connections or the compounding effect of professional and political connections. Finally, we investigate whether the effect of professional connections differs between firms with and without political connections. The results, as reported in panel C of Table 4, show that professional connections have a greater and stronger positive effect on trade credit in firms without political connections. This finding is consistent with our main argument that managerial professional connections have a significantly positive effect on firms' access to trade credit when we remove the potential compounding effect from political connections. 4.3. How do managerial professional connections help firms to obtain trade credit? We have provided evidence that managerial professional connections do play an important role in helping firms' access to trade credit by establishing reputation and interpersonal trust and thus reduce the credit risk faced by suppliers of trade credit. In this section, we aim to provide direct evidence to support our above argument. Following our discussion in Section 2.3, we first investigate the effect of managerial professional connections on trade credit in regions with different levels of social trust. In order to do so, we use the results in the survey conducted by the “Chinese Enterprise Survey System” in 2001 as a measure of regional social trust (the measure is also used by Wu et al. (2014)) to interact with our key variable of managerial professional connections. As can be seen from our results in Table 5, managerial professional connections and social trust are both statistically significantly positively related to trade credit, and the positive relationship between social trust and trade credit is consistent with Wu et al. (2014). The coefficient of the interaction term is negative and statistically significant at the 5% level of significance in the full sample (column 1) and the non-SOE subsample (column 3), suggesting a substitution effect of managerial professional connections and regional social trust. This result supports our expectation that managerial professional connections play a more important role when the regional social trust is lower (and the credit risk is higher). In order to provide further empirical evidence for our hypothesis H3b, we investigate how the relationship between professional connections and firms' access to trade credit is influenced by the business environment. We adopt the following measure to proxy the business

192

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

environment: the Herfindahl-Hirschman Index (HHI) of concentration to measure firms' competitive environment (Herfindahl, 1950; Gaspar and Massa, 2006). Please note that a higher HHI means an industry has more concentration and less competition. The results, as reported below in Table 6, show that HHI is significantly negatively correlated to trade credit, suggesting that firms in industries with higher market concentration (high HHI also means low market competition) obtain less trade credit. In addition, such high market concentration also reduces the positive relationship between managers' professional connections through industry associations and firms' access to trade credit (the coefficient of the interaction term PROCON*HHI is significantly negative). This result indicates that managers' professional connections are more important when firms face a highly competitive product market. Overall, our results in Tables 5 and6confirm our hypotheses H3a and H3b, and suggest that managerial professional connections play a more important role in helping firms to access trade credit when firms face greater difficulty in accessing trade credit (the credit risk is higher for the trade credit suppliers). 4.4. Alternative measurements of key variables 4.4.1. (Directors') professional connections and the financing component of trade credit In the above analysis, we use the working experience of firms' top executives – CEO/chairman – as a measurement of firms' professional connections and total account payable/total assets as a measurement of trade credit. One related question is: whether our results change if we use alternative proxies to measure our key variables. In order to address this issue, (1) we further use directors' professional connections as an alternative measure of firms' professional connections; (2) following Ge and Qiu (2007), we divide accounts payable into transaction and financing components, and using the financing component only as an alternative measure of trade credit, and we then conduct new regressions using those alternative measures. In order to do so, we create two new variables: (1) we define a new variable (DIR_PROCON), which is a dummy that equals 1 if at least one of the directors has professional connections and 0 otherwise (please note top executives – CEO/chairman – in Chinese firms are also board directors, so this definition of directors' professional connections includes both top executives' and other directors' professional connections); and (2), following Ge and Qiu (2007), we define a dependent variable (FIN_CREDIT) as total long-term trade credit/total assets to measure firms' trade credit for financing purposes. Our results, as reported in Table 7, show clearly that our findings still hold when we use alternative proxies to measure our key dependent and independent variables. 4.4.2. Professional connections and abnormal trade credit More recent literature documents that firms' trade credit is influenced by the nature of transacted goods, such as differentiated products, services and standard products (Fabbri and Menichini, 2010; Giannetti et al., 2011). By including the industry dummy and degree of competition variables in our regressions models, we have somehow controlled for the difference in the nature of goods. However, the nature of goods may still be a factor for firms in the same industry. Therefore, we further calculate the normal trade credit of each firm that may be needed in its day-to-day operations and obtain the residual as the abnormal trade credit it could take. By using the abnormal part of trade credit, we can differentiate the normal trade credit (associated with the nature of goods) and the abnormal trade credit it could take.8 By using the abnormal trade credit (ABNORMAL_CREDIT) as a new dependent variable in our baseline regression model, we are able to examine whether firms with professional connections also use more abnormal trade credit than firms without such connections. The results, as reported in Table 8, show a significantly positive relationship between managerial professional connections and the level of abnormal trade credit that firms use, and this relationship is found to be stronger in non-SOEs than in SOEs, a result 8 In particular, we use the following procedure to calculate the normal and abnormal level of trade credit: (1) We follow the approach by Banerjee et al. (2000) and Loof (2004) to estimate firms' optimal capital structure to each year and industry using the following equation:

Levi:t ¼ a0 þ a1 Sizei:tþ a2 Growi:t þ a3 Roei:t þ a4 Mogagei:t þ a5 Riski:t þ: a6 Ndtsi:t þ a7 Uniquei:t þ a8 Lqudtyi:t þ a9 Divdt i:t þ a10 Fdit i:t þ a j S j Industryi:t þ mi:t where Size is the firm's size, defined as the logarithm of firm's total asset value. Roe is the firm's return on equity. Grow is used to measure growth, which is defined as the growth rate of revenue. Mogage is the proxy for collateral, measured by the ratio of fixed assets plus inventory to total assets. Risk measures the firm's specific operational risk, calculated as the deviation of operation margin. Ndts is a measure of non-debt tax shield, defined as the proportion of accumulated depreciation to total assets. Variable Unique controls for the nature of goods, measured by the percentage of operation cost plus management fee to total revenue. Lqudty is the liquidity of assets, measured by the liquid ratio. Divdt is a dummy variable, which takes the value of 1 if the firm is cashing dividend in period t. Fdit is the firm's demand for cash; equal to dividend payment plus capital expenditure and net increase in operational capital, minus operational cash flow after tax. The model also controls for industry. (2) Obtain the firm's optimal debt. By estimating Eq. (A1), we can predict the optimal value of the capital structure ratio Lev_optimal, and obtain the optimal debt volume Debt_optimal, which equals Asset×Lev_optimal. (3) Calculate the firm's abnormal trade credit. We obtain the firm's normal trade credit by the following Eq. (A2): TC optimali:t ¼ Debt optimali:t −Debt indi:t where TC_optimal is Debt_optimal adjusted by industry median Debt_ind. The abnormal trade credit (TC_a) is calculated as follows: TC ai:t ¼ TC optimali:t −TC i:t ;

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

193

that is consistent with our prior results. In particular, we show that a 1% increase in professional connections results in a 0.989% increase of abnormal trade credit in all Chinese firms, and the increase is even higher in non-SOEs (1.072%). 4.5. Heterogeneity of professional connections and its impact on trade credit In our above analysis, we assume all types of professional connections have the same effect on trade credit. But in reality, managerial professional connections may not be homogenous. Therefore, in this section, we aim to provide further evidence to support our main argument by investigating the heterogeneity of managerial professional connections and its impact on firms' access to trade credit. 4.5.1. The effect of different levels of professional connections on firms' access to trade credit The industry associations in China vary at different levels: central level, provincial level and city level. Three new dummy variables (PROCON_CENTRAL, PROCON_PROVINCE and PROCON_CITY) are defined as 1 if a firm is professionally connected at the central level, province level or city level and 0 otherwise. Managers who have working experience with a higher level of industry associations are usually able to establish a stronger personal reputation and interpersonal trust and thus may play a more important role in helping firms' access to trade credit. Table 9 reports the results regarding different levels of social connections on firms' access to trade credit. The results confirm our expectation that the higher the professional connections, the stronger the influence on trade credit. 4.5.2. Current versus past working experience in an industry association If, as expected, working experience in an industry association helps the top executives to establish reputation and capture trust from other companies, it is reasonable to further conjecture that current working experience should play a more important role than past working experience in an industry association, because the longer it is since a top executive left his position in an

Table 9 The effect of different levels of professional connections on trade credit. This Table presents the effect that different levels of professional connections have on firms' access to trade credit. The dependent variable CREDIT is the total accounts payable to total assets; PROCON_CENTRAL is a dummy equal to 1 if the manager has a connection at the central level; PROCON_PROVINCE is a dummy equal to 1 if the manager has a connection at the provincial level; PROCON_CITY is a dummy equal to 1 if the manager has a connection at the city level. Coefficients for control variables are not reported to save space. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. Panel A full sample VARIABLES

CREDIT

PROCON_CENTRAL

0.015** (2.29)

ALL

ALL

PROCON_PROVINCE

ALL

0.014** (1.96)

PROCON_CITY Control variables YEAR INDUSTRY Observations Adjusted R-squared F

0.033 (1.47) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions 3986 3986 0.57 0.53 148.6*** 148.5***

3986 0.52 148.8***

Panel B SOE and non-SOE VARIABLES

PROCON_CENTRAL PROCON_PROVINCE PROCON_CITY Control variables YEAR INDUSTRY Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted-R2 F

Credit2 SOE

NSOE

0.0140 (1.358)

0.0113** (2.357)

SOE

NSOE

0.0017 (1.427)

0.0113** (2.357)

SOE

0.0107 (1.101) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions χ(2) = 18.62 χ(2) = 19.32 χ(2) = 26.05 2570 1416 2570 1416 2570 0.414 0.469 0.414 0.469 0.414 50.23 36.89 50.97 36.89 50.26

NSOE

0.0234 (1.604)

1416 0.470 37.06

194

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 10 Current vs past professional connections and trade credit. The aim of this Table is to test whether our results hold when we give current professional connections a higher weighting than past professional connections. We define a new variable (CURRENT_PROCON) as an alternative measure of professional connections to give 2 to firms whose top executives have current professional connections, 1 to firms whose top executives have past professional connections, and 0 to firms whose top executives do not have any professional connections. The dependent variable CREDIT is the total accounts payable to total assets. Coefficients for control variables are not reported to save space. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT ALL

CURRENT_PROCON Control variables YEAR INDUSTRY Test for difference of the coefficient of CURRENT_PROCON in SOE and NON-SOE Observations R-squared F

NON-SOE

SOE

0.017** 0.060*** 0.009 (2.345) (2.514) (0.493) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions χ(2) = 44.79 3986 2579 1407 0.104 0.138 0.144 13.14** 11.90** 16.625**

association, the harder it is for him to retain the trust he previously enjoyed. But in our previous analysis, we did not consider this issue because we gave current and past working experience the same weight. Thus we address this issue in this subsection. We first construct a variable (CURRENT_PROCON) to assign 2 to firms whose top executives have current professional connections, 1 to firms whose executives have past professional connections, and 0 to firms whose top executives do not have any professional connections, and use this new variable as a key independent variable and re-run our regression. We report our results below in Table 10. We find our results hold and the coefficients of our new variable are significant in both the full sample and the non-SOE subsample regressions, and, more importantly, the coefficients of CURRENT_PROCON are greater than those in Table 3, confirming that if we weight current connections more than past connections, professional connections have a greater impact on trade credit. 4.6. The endogeneity issue It is possible that managers' professional connections through industry associations are not exogenous, which means that our results may be influenced by a potential endogeneity issue caused by either omitted variables or reverse causality problems.9 The omitted variable problem means that professional connections and their impact on trade credit may be driven by omitted variables. For example, one omitted variable could be that firms with professional connections are those with higher growth or better performance. And thus they are able to obtain more trade credit. Reverse causality problems refer to the probability that it is not clear whether it is professional connections that result in more trade credit or whether more trade credit enables managers to form professional connections. This section aims to address the endogeneity issue using different approaches. For the omitted variables problem we use (1) the difference-in-difference method to examine the change in trade credit in firms with and without professional connections, under the shock of the global financial crisis; (2) the 2SLS model with instrument variables; and (3) the propensity score matching method to compare the ability of firms with and without professional connections to obtain trade credit. Then we compare the trade credit before and after firms acquire or lose professional connections to address the potential problems caused by reverse causality. The results are presented and discussed in this subsection. 4.6.1. Addressing the endogeneity issue caused by omitted variables The global financial crisis has an important influence on corporate financial decisions because it causes a sudden failure of the credit market. Due to concerns about higher credit risk during the crisis period, firms tend to provide less trade credit to accelerate cash conversion, so credit-receiving firms may find it harder to access trade credit during the crisis period. We expect that professional connections should play a more important role during the crisis period because they effectively mitigate the credit risk between demanders and suppliers of trade credit and thus reduce the inverse effect that the crisis has on firms' access to trade credit. Thus, compared to those without professional connections, firms with connections should be less influenced by the financial crisis. As the global financial crisis hit the Chinese economy in August 2008, and its influence lasted for about one year to 2009, we create two dummy variables, DUM08 and DUM09 to capture its influence. We then add the two dummy variables into our baseline regression and interact them with firms' professional connections. The results are reported in Table 11. As expected, our results show that coefficients of the two dummies are both significantly negative, which suggests that the crisis has an inverse effect

9 The endogeneity issue could also be caused by a measurement error in key variables, but we have addressed this issue from our results in Section 4.4 by using alternative proxies to measure our key variables.

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

195

Table 11 The effect of the global financial crisis on the relationship between professional connections and firms' access to trade credit. This Table presents the effect that managerial professional connections and the global financial crisis have on firms' ability to obtain trade credit. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections; DUM08 is a dummy equal to 1 for all observations in the year 2008 and 0 for all other observations; DUM09 is a dummy equal to 1 for all observations in the year 2009 and 0 for all other observations. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT ALL

SOE

NON-SOE

ALL

SOE

NON-SOE

PROCON

0.012** (2.369) −0.061*** (−16.312) 0.053*** (4.571)

0.010 (1.407) −0.037 (−1.189) 0.077 (0.401)

0.021*** (2.687) −0.097 (−0.923) 0.014*** (2.456)

0.013* (1.751)

0.001 (1.003)

0.051*** (5.144)

DUM08 DUM08*PROCON DUM09 DUM09*PROCON Control variables YEAR INDUSTRY P-value for interaction Test for difference of the coefficient of CURRENT_PROCON in SOE and NON-SOE Observations Adjusted–R2 F

−0.038*** −0.051 (−9.914) (−0.352) 0.007* 0.005 (1.77) (1.485) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions P = 0.027 P = 0.0426 P = 0.0012 P = 0.001 P = 0.0231 χ(2) = 46.18 χ(2) = 32.47

−0.033*** (−2.716) 0.031*** (3.086)

3986 0.60 166.5**

1416 0.474 37.03*

2570 0.414 49.56**

1416 0.472 36.81**

3986 0.593 69.14*

2570 0.414 49.64***

P = 0.000

on firms' access to trade credit. However, the interaction terms of DUM_2008*PROCON and DUM_2009*PROCON are both significantly positive, indicating that the effect the crisis has on firms' access to trade credit is weakened by professional connections. We further conduct the Heckman two-stage estimation with instrument variables to address the endogeneity of our key independent variable – professional connections. This study selects two types of instrument variables that may have an influence on the likelihood of managerial professional connections: (1) the number of employees of the firm; and (2) the educational background of the managers. The number of employees is expected to have a positive effect on the likelihood of a firm's top executives becoming professionally connected. This is because in China having a large number of employees is always important for the stability of the economy, so managers of firms with more employees are more likely to be selected as executives of an industry association. We further expect that better educated managers may be more likely to be selected as executive members of industry associations, so we choose managerial education as another instrument variable. To do so, we create the following two instrument variables: (1) The number of employees (NUMEMPLOYEE) measured by the natural logarithm of the total number of employees at the end of the year; (2) CEO education (CEOEDU) measured by the educational background of the company's CEO/Chairman. The variable equals 6 for doctor and above; 5 for master's degree; 4 for bachelor degree; 3 for college and diploma degree; 2 for school degree; and 1 for other lower levels of educational background. The key to the Heckman two-stage regression is that the instrument variables should (should not) have an influence on the probability of managerial professional connections (firms' access to trade credit). Thus for validation of the our instrument variables, we first test whether those instrument variables have an influence on our dependent variables – firms' access to trade credit. The results, as reported in column 1 of Table 12, show that the regression coefficients for our instrument variables are all insignificant when we regress trade credit on those instrument variables only, suggesting that our selection of instrument variables is valid. We report our first stage regression results in column 2 of Table 12. As expected, the instrument variables, the number of employees and the educational background of top executives, both have a statistically significantly positive effect on the probability of firms being professionally connected. The second stage regression results, as reported in columns 3 to 5 of Table 12,further confirm that our main results hold when we address the potential endogeneity issue using the Heckman two-stage approach. One further concern about our results is that firms with and without professional connections vary in firm-specific characteristics, which means that our findings may be caused by the different firm-specific characteristics rather than the professional connections. We therefore compare the trade credit in firms with and without professional connections using the propensity score matching method based on matching variables. In the results reported in Table 13, we match each firm with professional connections to a firm without such connections. The match criteria are, firm size for model 1; firm performance for model 2; number of employees for model 3; product market share for model 4; firm size, performance, number of employees, and market share for model 5; firm size, performance, number of employees, market share and all control variables including year and industry in regressions for model 6. Our results support our

196

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 12 Heckman two-stage regression results with instrument variables. This Table reports the Heckman two-stage regressions on the effect of professional connections on trade credit. We choose two instrument variables: NUMEMPLOYEE is the natural logarithm of total number of employees at the year end. CEOEDU is the measure of the educational background of the CEO/chairman. PROCON^ is the predicted value of managerial professional connections estimated from the first stage regression. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

NUMEMPLOYEE CEOEDU

CREDIT

0.041 (1.601) 0.001 (1.383)

First stage

Second-stage

PROCON

CREDIT SOE

NON-SOE

0.016** (2.284) 0.014*** (2.512) 0.026*** (5.642) 0.004*** (2.033) −0.174* (−1.780) 0.014 (0.960) 0.010 (0.347) 0.014 (1.393) 0.039*** (4.412) 0.279*** (2.595)

0.019 (1.153) 0.060 (0.307) 0.059 (1.208) 0.002*** (2.292) −0.080** (−2.214) 0.080*** (5.194) 0.034 (1.118) 0.061 (1.003) 0.068*** (7.551) 0.153*** (3.337)

0.012* (1.982) 0.012*** (3.125) 0.016** (2.273) 0.011*** (2.672) −0.007 (−0.327) 0.024 (1.138) 0.057 (1.401) 0.021 (1.249) 0.091*** (6.478) −0.151** (2.390)

2570 0.368

1416 0.46

46.33***

28.63*

0.015** (2.081) 0.002*** (2.582)

PROCON^ SIZE

ALL

0.046 (0.730) 0.023* (1.673) 0.210 (0.617) −0.107*** (−2.487) 0.060 (1.317) 0.140 (0.349) 0.086 (0.797) 0.013 (0.375) 0.040 (0.408) Controlled in all regressions Controlled in all regressions 3986 3986 0.235 0.29 38.14 0.001 32.86** 47.35**

AGE EBIT FIXASSET OPCASH ROA PC LEV CONSTANT YEAR INDUSTRY Observations Adj-R2 Weak IV F Statistics Hansen's J Statistics F

main argument that firms' access to trade credit differs significantly between firms with and without professional connections through industry associations even with control for various different factors. 4.6.2. Addressing the endogeneity issue caused by reverse causality In this subsection, we address the reverse causality problem by observing whether firms' trade credit changes after they acquire or lose their professional connections. As firms' acquisition and loss of managerial professional connections mostly happened in four particular years, 2007, 2008, 2011 and 2012 (some firms acquired social connections in 2007 and 2008, while some others lost professional connections in 2011 and 2012), we focus our tests on the above four years. Interestingly, the results in Table 14 suggest that firms increase the use of trade credit after a manager with professional connections is acquired in 2007 and 2008, which confirms our main argument that managerial professional connections help firms to obtain trade credit. While firms' access to trade credit decreases after firms lose managers with professional connections, the insignificant result suggests that the influence of managerial professional connections does not disappear immediately, and that there is some residual effect after firms lose the professional connections. Table 13 Comparison of firms' access to trade credit in firms with and without professional connections. We match each firm with professional connections to a firm without such connections. The match criteria are: firm size for model 1; firm performance for model 2; number of employees for model 3; product market share for model 4; firm size, performance, number of employees, and market share for model 5; firm size, performance, number of employees, market share and all control variables including year and industry in regressions for model 6. The bootstrap and sampling without replacement (50 times) method is used. Bootstrap standard errors, Z-value and P-value are reported below the coefficients for all models. *, **, *** represent significant at the 10%, 5% and 1% level significance respectively.

Coefficient Bootstrapstandard errors Z-value P-value

(1)

(2)

(3)

(4)

(5)

(6)

0.0105* 0.004 1.98 0.048

0.0106** 0.005 2.07 0.039

0.0103* 0.005 1.93 0.054

0.0124** 0.005 2.26 0.0242

0.0125* 0.006 1.77 0.076

0.0129** 0.006 2.25 0.025

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

197

Table 14 Firms' access to trade credit before and after the professional connections are acquired or lost. This Table compares the trade credit before and after the status of managers' professional connections changes. We observe that some firms acquire professional connections in 2007 and 2008, while some other firms lose professional connections in 2011 and 2012. There is very little change in professional connections in other years. So we conduct our analysis based on the above four years. *, ** and *** represent significance at the 10%, 5% and 1% levels. Time

After acquire

Before acquire

Difference

2007 P-value 2008 P-value

0.107*** 0.000 0.087*** 0.000

0.081*** 0.000 0.071*** 0.000

0.027** 0.0103 0.017* 0.068

2011 P-value 2012 P-value

Before lose

After lose

Difference

0.111*** 0.000 0.100*** 0.000

0.099*** 0.000 0.08*** 0.000

0.012 0.333 0.02 0.257

4.7. Robustness results In this section, we aim to provide additional evidence for our main story. In particular, the following four questions will be addressed: (1) whether firms with professional connections also offer more trade credit to other firms; (2) if yes, whether the additional trade credit brought by managers' professional connections is more than the additional trade credit that they offer to other firms; (3) whether and how managerial professional and political connections have an influence on firms' access to formal financing resources, such as bank credit; (4) whether our results are influenced by the large size effect. To answer these questions, we conduct the further following tests. 4.7.1. The effect of professional connections on supply of trade credit and the net position of trade credit In Table 15, we first use firms' supply of trade credit, measured by total accounts receivable to total assets, as a dependent variable (S_CREDIT) to examine whether managers' professional connections also have a positive effect on supply of trade credit. The results in column 1 show that the coefficient for managers' professional connections is positive (0.015) and is statistically significant at the 5% level of significance. The result indicates that managers also have a strong incentive to offer trade credit to their customers when they have professional connections through industry associations. By doing so, they can maintain their industry reputation and receive more trade credit. In column 4, we use the net position of trade credit (NET_CREDIT), measured by the difference between accounts payable and accounts receivable to total assets as a dependent variable. If this variable is more than 0, it means that accounts payable are more than accounts receivable. And similarly, if the estimated coefficient for managers' professional connections is positive, that means managers' professional connections help firms to receive more trade credit than they offer. Our results show a significantly positive relationship between managers' professional connections and the net position of trade credit that firms receive. 4.7.2. The effect of professional connections and political connections on bank credit We have provided evidence that managerial connections play an important role in helping firms' access to informal financial resources – trade credit – while political connections do not. One additional question people may ask is: what about formal financial resources – bank credit? Do professional connections also have an influence on bank credit? In order to answer this question, we conduct further tests on the effect of professional and political connections on bank credit. Table 16 reports the results. Table 15 The effect of managers' professional connections on supply of trade credit, net trade credit. The dependent variable is firms' supply of trade credit (S_CREDIT) for model 1, net positive trade credit (NET_CREDIT) for model 2.PROCON is a dummy for managerial professional connections with industry associations; ALL/SOE/NON-SOE refer to regressions using the full sample/SOE sample/non-SOE sample separately. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

S_CREDIT ALL

PROCON Control variables YEAR INDUSTRY Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted–R2 F

NET_CREDIT SOE

NON-SOE

ALL

SOE

0.015** 0.0259 0.011*** 0.022** 0.030 (2.29) (0.316) (2.766) (2.35) (0.612) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions χ(2) = 37.64 χ(2) = 44.85 3986 2570 1416 3986 2570 0.57 0.353 0.365 0.43 0.376 148.6*** 87.02* 121.42*** 85.85** 142.7**

NON-SOE 0.0198*** (5.517)

1416 0.347 150.05**

198

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Table 16 The effect of professional connections through industry associations and political connections on bank credit. The dependent variable is firms' bank loan ratio (BANKLOAN)·PROCON is a dummy for managers' professional connections; PC is a dummy for managers' political connections. Detailed definitions for all variables are reported in Appendix A. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. Standard errors are clustered by company. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

ALL

BANKLOAN SOE

NON-SOE

PROCON

0.002 (1.54)

0.014 (1.61)

0.025 (1.29)

PC Control variables SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions INDUSTRY Controlled in all regressions YEAR Controlled in all regressions Test for difference of the coefficient of CURRENT_PROCON in SOE and NON-SOE χ(2) = 36.18 Observations 3986 2579 1407 0.416 0.404 0.445 Adjusted-R2 F 187.7 55.80 63.20

ALL

SOE

NON-SOE

0.018* (1.94)

0.011 (1.14)

0.014*** (2.61)

3986 0.416 187.6

χ(2) = 104.71 1407 2579 0.443 0.404 62.55 55.91

Our results show that, although managers' professional connections show a positive effect on firms' access to bank credit, the coefficients are statistically insignificant, no matter whether we conduct our analysis using the full sample, the SOE sample or the non-SOE sample. However, we do find a significantly positive relationship between political connections and firms' access to bank credit, a result that is consistent with previous studies, such as Li et al. (2008) and Firth et al. (2009). These results complement our story by showing that political connections play a more important role when firms access bank credit, while professional connections are relatively more important when firms access trade credit. 4.7.3. Professional connections and trade credit in firms with different size Our results have shown that large firms tend to have more trade credit, but it is not known whether our results are driven by the large firm effect, which means that managers of large firms may be more likely to build professional connections and so the positive relationship between professional connections and trade credit may only exist in large firms. Thus in this section, we separate our sample into two subsamples (large and small firms) according to firm size and then repeat our regressions on the relationship between professional connections and trade credit in the two subsamples. As shown in Table 17, we find that managerial professional connections have a significantly positive effect on firms' access to trade credit in both the large and small-sized subsamples, suggesting that our main results are not driven by the large size effect. 5. Conclusions This study examines the effect of managers' professional connections and political connections on firms' access to trade credit, using a sample of Chinese listed firms. We find that managerial professional connections with industry associations help firms, especially non-SOEs, to obtain trade credit, while political connections do not. We further document that managerial professional connections play a more important role in firms that locate in regions with low social trust and firms that operate in industries with higher product market competition. Our findings hold when we use directors' professional connections as an alternative measure of professional connections, and when we measure trade credit using the financing component of trade credit and

Table 17 The effect of professional connections on firms' access to trade credit in firms of different size. This Table presents the effect that professional connections have on firms' access to trade credit in small and large-sized firms. ALL/SOE/NON-SOE refer to regressions using the full sample/SOE sample/non-SOE sample separately. The dependent variable CREDIT is the total accounts payable to total assets; PROCON is a dummy for managers' professional connections. Detailed definitions for all variables are reported in Appendix A. Standard errors are clustered by company. The sample is trimmed at 1% on each variable in each tail. T-statistics are reported in brackets below the coefficients. *, ** and *** represent significance at the 10%, 5% and 1% levels. VARIABLES

CREDIT Small size ALL

PROCON Control variables INDUSTRY YEAR Test for difference of the coefficient of PROCON in SOE and NON-SOE Observations Adjusted-R2 F

Large size SOE

NSOE

ALL

SOE

0.023*** 0.014 0.036* 0.012* 0.004 (3.622) (1.240) (1.890) (1.722) (0.211) PC, SIZE, AGE, FIXASSET, EBIT, ROA, LEV included in all regressions Controlled in all regressions Controlled in all regressions χ(2) = 23.18 χ(2) = 16.97 1993 1158 835 1993 1412 0.349 0.340 0.379 0.385 0.345 120.3*** 47.31** 34.61* 166.7** 18.71**

NSOE 0.099*** (5.136)

581 0.359 24.04**

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

199

abnormal trade credit. Moreover, we show that, compared to a lower level of professional connections and past professional connections, firms with a higher level of professional connections and current connections use significantly more trade credit. Using the global financial crisis as a natural experiment, we find that managers' professional connections play a more important role in helping firms to obtain trade credit in the financial crisis period from 2008 to 2009. We provide evidence that firms increase their use of trade credit after they acquire new social connections. Our results are found to be robust when we use both the Heckman two-stage approach and the propensity score matching method. Moreover, we document that, although firms with professional connections also offer more trade credit to other firms, they still obtain more trade credit after the trade credit they offer to other firms is deducted, and thus confirm our main expectation that managers' professional connections help firms to access trade credit. Our study provides evidence that in emerging markets with poor protection of creditors and a poorly developed formal financing market, firms, especially non-SOEs, which have limited access to the formal financing market tend to rely on trade credit as an important financing channel. Under these circumstances, managerial professional connections play an important role in helping firms to access trade credit by establishing reputation and trust, especially to firms which face greater credit risk, i.e., firms located in regions with lower social trust, firms operating in highly competitive industries and during the global financing crisis period. Overall, we argue that managerial professional connections, rather than political connections, help firms, especially those with limited access to formal financing, to obtain informal financing resources. Appendix A. Definition of variables Variable name

Detailed definition

Dependent variables CREDIT FIN_CREDIT ABNORMAL_CREDIT S_CREDIT NET_CREDIT

The ratio of accounts payable to total assets Financing component of trade credit, total long-term trade credit/total assets See footnote 7 for the calculation of abnormal trade credit The ratio of accounts receivable to total assets (Accounts payable minus accounts receivable)/total assets

Key independent variables PROCON A dummy variable equal to 1 if the general manager/CEO or chairman of the firm has working experience or is currently an executive member, such as secretary, associate secretary, chairman, associate chairman, director or executive director, of an industry association, and 0 otherwise. DIR_PROCON A dummy variable equal to 1 if at least one of the directors has working experience or is currently an executive member, such as secretary, associate secretary, chairman, associate chairman, director or executive director, of an industry association, and 0 otherwise. PC A dummy variable equal to 1 if the chairman or general manager/CEO currently serves or formerly served in the government or military, or serves/served as a deputy of the National/Provincial People's Congress or the People's Political Consultative Conference, and 0 otherwise. Other variables TRUST HHI BANKLOAN SIZE AGE FIXASSET LEVERAGE OPCASH EBIT ROA NUMEMPLOYEE CEOEDU

The regional level of social trust following the survey conducted by the “Chinese Enterprise Survey System” in 2001 The Herfindahl-Hirschman Index (HHI) of concentration to measure firms' competitive environment Total bank loans to total assets Natural logarithm of total assets Observation year minus the year in which the firm is founded Total tangible assets to total assets Total debt to total assets Total operating cash flow to total assets Total EBIT (earnings before tax and interest) to total assets Total net earnings to total assets Natural logarithm of the total number of employees at the year end Educational background of the company's CEO/Chairman. The variable equals 6 for doctor and above; 5 for master's degree; 4 for bachelor degree; 3 for college and diploma degree; 2 for school degree and 1 for other lower levels of educational background.

References Adhikari, A., Derashid, C., Zhang, H., 2006. Public policy, political connections, and effective tax rates: longitudinal evidence from Malaysia. J. Account. Public Policy 25, 574–595. Agrawal, A., Knoeber, C.K., 2001. Do some outside directors play a political role? J. Law Econ. 44, 179–198. Alphonse, P., Ducret, J., Severin, E., 2003. When Trade Credit Facilitates Access to Bank Finance: Evidence from US Small Business Data. University of Lille, Working Paper. Atanasova, C., Wilson, N., 2004. Disequilibrium in the UK corporate loan market. J. Bank. Financ. 28, 595–614. Ayyagari, M., Demirg¨uc¸-Kunt, A., Maksimovic, V., 2010. Formal versus informal finance: evidence from China. Rev. Financ. Stud. 23, 3048–3097. Berkman, H., Cole, R.A., Fu, L.J., 2010. Political connections and minority shareholder protection: evidence from securities market regulation in China. J. Financ. Quant. Anal. 45 (6), 1391–1417. Biais, B., Gollier, C., 1997. Trade credit and credit rationing. Rev. Financ. Stud. 10, 903–937. Bougheas, S., Mateut, S., Mizen, P., 2009. Corporate trade credit and inventories: new evidence of a trade-off from accounts payable and receivable. J. Bank. Financ. 33, 300–307. Brennan, M., Maksimovic, V., Zechner, J., 1988. Vendor financing. J. Financ. 43, 1127–1141.

200

Q. Liu et al. / Journal of Corporate Finance 41 (2016) 179–200

Burkart, M., Ellingsen, T., 2004. In-kind finance: a theory of trade credit. Am. Econ. Rev. 94, 569–590. Chen, C.J.P., Li, Z., Su, X., Sun, Z., 2011. Rent-seeking incentives, corporate political connections, and the control structure of private firms: Chinese evidence. J. Corp. Fin. 17, 229–243. Claessens, S., Feijen, E., Laeven, L., 2008. Political connections and preferential access to finance: the role of campaign contributions. J. Financ. Econ. 88, 554–580. Cull, R., Xu, L., 2003. Who gets credit? The behaviour of bureaucrats and state banks in allocating credit to Chinese state-owned enterprises. J. Dev. Econ. 71 (2), 533–559. Cull, R., Xu, L.C., Zhu, T., 2009. Formal finance and trade credit during China's transition. J. Financ. Intermed. 18, 173–192. Cunat, V., 2007. Trade credit: suppliers as debt collectors and insurance providers. Rev. Financ. Stud. 20, 491–527. Cunningham, R., 2004. Trade credit and credit rationing in Canadian firms. Working Paper 2004–49. Canada, Bank of. Daripa, A., Nilsen, J., 2005. Subsidizing inventory: a theory of trade credit and repayment. Working Papers in Economics and Finance No. 0522. Birkbeck College. Demirguc-Kunt, A., Maksimovic, V., 2002. Firms as Financial Intermediaries: Evidence from Trade Credit Data. University of Maryland, Working Paper. Fabbri, D., Menichini, A.M., 2010. Trade credit, collateral liquidation, and borrowing constraints. J. Financ. Econ. 96, 413–432. Faccio, M., 2006. Politically connected firms. Am. Econ. Rev. 96 (1), 369–386. Faccio, M., Masulis, R., McConnell, J.J., 2006. Political connections and corporate bailouts. J. Financ. 61, 2597–2635. Firth, M., Lin, C., Liu, P., Wong, S.M.L., 2009. Inside the black box: bank credit allocation in China's private sector. J. Bank. Financ. 33 (6), 1144–1155. Fisman, R., Love, I., 2003. Trade credit, financial intermediary development and industry growth. J. Financ. 58, 353–374. Frank, M., Maksimovic, V., 2005. Trade Credit, Collateral, and Adverse Selection. University of Maryland, Working Paper. Gaspar, J.S., Massa, M., 2006. Idiosyncratic volatility and product market competition. J. Bus. 79, 3125–3152. Ge, Y., Qiu, J., 2007. Financial development, bank discrimination and trade credit. J. Bank. Financ. 31, 513–530. Giannetti, M., 2003. Do better institutions mitigate agency problems? Evidence from corporate finance choices. J. Financ. Quant. Anal. 38, 185–212. Giannetti, M., Burkart, M., Ellingsen, T., 2011. What you sell is what you lend? Explaining trade credit contracts. Rev. Financ. Stud. 24, 1261–1298. Gu, F., Hung, K., Tse, D.K., 2008. When does guanxi matter? Issues of capitalization and its dark sides. J. Mark. 70, 12–28. Guariglia, A., Mateut, S., 2006. Credit channel, trade credit channel, and inventory investment: evidence from a panel of UK firms. J. Bank. Financ. 30, 2835–2856. Guiso, L., Sapienza, P., Zingales, L., 2004. The role of social capital in financial development. Am. Econ. Rev. 94, 526–556. Haselmann, R., Schoenherry, D., Vig., V., 2013. Lending in Social Networks. University of Warwick working paper. Herfindahl, O.C., 1950. Concentration in the Steel Industry. Columbia University, PhD dissertation. Hsiung, B., 2013. Guanxi: personal connections in Chinese society. J. Bioecon. 15, 17–40. Jain, N., 2001. Monitoring costs and trade credit. Q. Rev. Econ. Finance 41, 89–110. Jha, A., Chen, Y., 2015. Audit fees and social capital. Account. Rev. 90, 611–639. Karlan, D.S., 2005. Using experimental economics to measure social capital and predict financial decisions. Am. Econ. Rev. 95, 1688–1699. Khwaja, A., Mian, A., 2005. Do lenders favor politically connected firms? Rent provision in an emerging financial market. Q. J. Econ. 120 (4), 1371–1411. Kohler, M., Britton, E., Yates, T., 2000. Trade Credit and the Monetary Transmission Mechanism. Working Paper, Bank of England. Le, N.T.B., Nguyen, T.V., 2009. The impact of networking on bank financing: the case of small and medium-sized enterprises in Vietnam. Enterp. Theory Pract. 33, 867–887. Lee, Y., Stove, J., 1993. Product risk, asymmetric information, and trade credit. J. Financ. Quant. Anal. 28, 285–300. Li, H., Meng, L., Wang, Q., Zhou, L., 2008. Political connections, financing and firm performance: evidence from Chinese private firms. J. Dev. Econ. 87, 283–299. Lin, T.T., Chou, J.H., 2014. Trade credit and bank loan: evidence from Chinese firms. Int. Rev. Econ. Financ. (forthcoming). Liu, Q., Tian, G., 2012. Controlling shareholder, expropriations and firm's leverage decision: evidence from Chinese non-tradable share reform. J. Corp. Fin. 18, 782–803. Liu, Q., Tang, J., Tian, G., 2013. Does political capital create value in the IPO market? Evidence from China. J. Corp. Fin. 23, 395–413. Liu, Q., Pan, X., Tian, G., 2016. To what extent did the economic stimulus package influence bank lending and corporate investment decisions? Evidence from China. J. Bank. Financ. (forthcoming). Long, M., Malitz, I., Ravid, A., 1993. Trade credit, quality guarantees and product marketability. Financ. Manag. 22, 117–127. Loof, H., 2004. Dynamic optimal capital structure and technical change. Struct. Chang. Econ. Dyn. 15, 449–468. Love, I., Preve, L., Sarria-Allende, V., 2005. Trade credit and bank credit: evidence from recent financial crises. Policy Research Working Paper 3716. Bank, World. Luk, C.L., Yau, O.H.M., Sin, L.Y.M., Tse, A.C.B., Chow, R.P.M., Lee, J.S.Y., 2008. The effects of social capital and organizational innovativeness in different institutional contexts. J. Int. Bus. Stud. 39 (4), 589–612. Mateut, S., Bougheas, S., Mizen, P., 2006. Trade credit, bank lending and monetary policy transmission. Eur. Econ. Rev. 50, 603–629. Mian, S., Smith, C., 1992. Accounts receivable management. J. Financ. 47, 169–200. Ng, C., Smith, J., Smith, R., 1999. Evidence on the determinants of credit terms in interfirm trade. J. Financ. 54, 1109–1129. Nilsen, J.H., 2002. Trade credit and the bank lending channel. J. Money Credit Bank. 34 (1), 226–253. Park, S.H., Luo, Y., 2001. Guanxi and organizational dynamic: organizational networking in Chinese firms. Strateg. Manag. J. 22 (5), 455–477. Peng, M.W., 2001. How entrepreneurs create wealth in transition economies. Acad. Manag. Exec. 15, 95–110. Peng, M.W., Luo, Y., 2000. Managerial ties and firm performance in a transition economy: the nature of micro-macro link. Acad. Manag. J. 43 (3), 486–501. Petersen, M., Rajan, R., 1997. Trade credit: theories and evidence. Rev. Financ. Stud. 10, 661–697. Preve, L., 2003. Financial Distress and Trade Credit: An Empirical Analysis. University of Texas, Working Paper. Rajan, R.G., Zingales, L., 1995. What do We know about capital structure? Some evidence from international data. J. Financ. 50 (5), 1421–1460. Smith, J., 1987. Trade credit and informational asymmetry. J. Financ. 42, 863–876. Wilner, B., 2000. The exploitation of relationships in financial distress: the case of trade credit. J. Financ. 55, 153–178. Wu, W., Firth, M., Rui, O.M., 2014. Trust and the provision of trade credit. J. Bank. Financ. 39, 146–159.