CHINA JOURNAL OF ACCOUNTING STUDIES 2018, VOL. 6, NO. 1, 63–83 https://doi.org/10.1080/21697213.2018.1483589 Does Foreign Bank Entry Aﬀect Domestic Enterprise Innovation?* a b c Jun Bai , Qingxi Meng and Yanyan Shen School of Economics and Management, Research Center of Corporate Governance and Management Innovation, Shihezi University, China; School of Accountancy, Shanghai University of Finance and Economics, China; School of Business, Renmin University of China, China ABSTRACT KEYWORDS banking competition; How does ﬁnancial opening-up impact the real economy? The diﬀerence-in-diﬀerence answer is important to China, the largest economy in transition approach; enterprise experiencing a critical period of further reform of the ﬁnancial system. innovation; foreign bank We capture the inﬂuence of ﬁnancial reform through foreign bank entry entry in this paper. Starting with a description of the process of foreign bank entry in China, we sum up two paths through which foreign bank entry aﬀects China’s domestic enterprise innovation: the direct eﬀect and the spillover eﬀect. Then, using the data of public listed companies in China from 2001 to 2014, we investigate the eﬀects of foreign bank entry on domestic enterprise innovation by exploiting the staggered multiple exogenous shocks and using a diﬀerence-in-diﬀerence approach. The results show that foreign bank entry contributes signiﬁcantly to domestic enterprise innovation, and they are robust when considering the potential concerns of omitted variables and reverse causality. Further study illustrates that the nurturing role derives not only from the direct eﬀect (although it is very limited), but also from the spillover eﬀect of foreign bank entry by enhancing competition among domestic banks. The ﬁndings suggest that it is necessary to mitigate the barriers to foreign banks and expand the banking sector in China. 1. Introduction Reform and opening-up in ﬁnance comes with the trend of economic globalisation. As an essential aspect of both in China, foreign bank entry calls for extensive attention (Lin, 2011). A growing number of emerging economies are admitting foreign banks in order to improve the eﬃciency of the domestic banking system with more capital, technology and advanced management available (Chen & Weng, 2013). Consequently, the question to be posed is how foreign bank entry impacts domestic real economies. Any answers provided will not only contribute to a reasonable evaluation about existing ﬁnancial policies, but also provide indicators for ﬁnancial reform in the future. Ample literature on the consequences of foreign bank entry has mainly focused on the bank-level, such as the impact on domestic banks’ operating eﬃciency and perfor- mance (Levy & Micco, 2007; Li & Han, 2008; Maria et al., 2004; Unite & Sullivan, 2003; CONTACT Qingxi Meng firstname.lastname@example.org; *Paper accepted by Guliang Tang. © 2018 Accounting Society of China 64 BAI ET AL. Zhang & Wu, 2010). However, the ultimate goal of ﬁnancial reform and opening-up is serving the real economy, so a discussion on how foreign bank entry aﬀects the real economy, and especially ﬁrms, provides a potential contribution. However, important disagreements remain among some scholars with respect to this issue. On the one hand, it is doubtful to what extent foreign bank entry could play a positive role in the domestic real economy, given the possibility of foreign banks tending to serve clients or overseas branch oﬃces from their parent countries even after their entry (Grosse & Goldberg, 1991; Yamori, 1998) and the inherent problem of information asymmetry (Stiglitz & Andrew, 1981). For example, Chen, Zong, and Zhang (2007), using survey data in China ﬁnd that among all clients of foreign banks, foreign-funded enterprises account for 70%, state-owned enterprises account for 20%, and private enterprises merely account for 10%. Taking Shanghai as an example, Tian and Xu (2010) also conclude that foreign bank entry contributes to foreign economic development, but it has an adverse impact on resource acquisition by domestic enterprises. On the other hand, some scholars argue that foreign bank entry does play a positive role on domestic enterprises. Berger, Klapper, and Udell (2001) conclude that foreign banks could overcome the geographical disadvantages and work more eﬃciently than domestic banks, hence helping the ﬁnancing of all enterprises on average. This view is further summed up as a ‘performance hypothesis’ by Degryse, Havrylchyk, and Jurzyk (2009). Furthermore, Chen and Weng (2013) show that foreign bank entry signiﬁcantly reduces the ﬁnancing cost of enterprises, with technology spillover acting as the main mechanism. Using foreign bank entry as a proxy for bank reform in China, Tsai, Chen, Lin, and Hung (2014) ﬁnd that the eﬃciency of capital allocation is improved after foreign bank entry, in terms of mitigating the politically-oriented investment problem for state-con- trolled listed companies and alleviating the problem of underinvestment in non-state- controlled ones. Notwithstanding this, the impact of foreign bank entry may vary with the heterogeneity of the enterprise. For instance, according to the ‘portfolio composition hypothesis’ raised by Degryse et al. (2009), when facing competition from foreign banks, domestic banks would shift their loan portfolio to informationally opaque borrowers, where informationally transparent ﬁrms would be able to take advantage of foreign bank entry. Lin (2011) further studies the heterogeneity and ﬁnds no signiﬁcant change on average in the amount of long-term bank loans among publicly-traded non-ﬁnancial ﬁrms in China after foreign bank entry. However, proﬁtable ﬁrms and non-stated ﬁrms get more long-term bank loans. Yao, Wu, and Wang (2015) show that foreign bank entry has little eﬀect on state-owned enterprises, and its impact on private enterprises varies over ﬁrm size. In large private enterprises, ﬁnancing constraints are released, while in small and medium-sized ones, they are signiﬁcantly tightened. As previously mentioned, there has been ample literature focusing on the direct eﬀect of foreign bank entry, while basically overlooking its spillover eﬀect of intensifying the competition and facilitating the management eﬃciency among domestic banks. With China’s banking system concentrated both in market structure and ownership, the spillover eﬀect of foreign bank entry might be more remarkable than the direct eﬀect. To recognise what will happen after foreign bank entry, we simultaneously take the direct eﬀect and the spillover eﬀect into account. To alleviate endogeneity concerns between banking competition and innovation, we exploit China’s accession to the World Trade Organization (WTO), a unique policy experiment, and employ a diﬀerence-in- CHINA JOURNAL OF ACCOUNTING STUDIES 65 diﬀerence (DID) approach to study the eﬀects of ﬁnancial opening on domestic econ- omy. China joined the WTO in 2001 and started to open up its banking sector gradually. According to the WTO negotiations, after a ﬁve-year transition period, ending in 2006, the banking sector in China would comprehensively open up. Since then, foreign banks have either set up branches or hold shares of domestic banks to enter China. The opening up of the banking sector is gradually expanding as a process. Furthermore, the choices of timing and region of foreign bank entry are virtually decided by the central government rather than the speciﬁc needs of local enterprises (Lin, 2011), thus consti- tuting a staggered multiple exogenous shock. Such an event oﬀers us a unique setting to avoid the ubiquity of endogeneity concerns between ﬁnance and economic growth, which further allows us to estimate accurately the eﬀect of foreign bank entry on the real economy in China. Unlike previous studies, we examine the economic output of foreign bank entry from the perspective of domestic enterprise innovation, for the following reasons. First, as the driving force of enterprises’ long-term development, innovation is very important. Second, enterprise innovation is often plagued by ﬁnan- cing constraints and corporate governance (Li, Gao, & Chen, 2015), so we could observe the improvement of the ﬁnancial environment brought by foreign bank entry and its function of liabilities governance at the same time. The paper starts by reviewing the process of foreign bank entry after China’s acces- sion to the WTO, and then theoretically identiﬁes the mechanism of how foreign bank entry aﬀects China’s domestic enterprise innovation, in terms of both direct and spillover eﬀect. We then examine the relationship between foreign bank entry and domestic enterprise innovation by utilising the staggered multiple exogenous shocks and a DID approach, using the data of Chinese listed companies. Our results show that foreign bank entry signiﬁcantly promotes the innovation of domestic enterprises. The results are robust when considering the potential problems of omitted variables and reverse causality. Further study illustrates that the nurturing role derives not only from the direct eﬀect (although it is very limited), but also from the spillover eﬀect of foreign bank entry by enhancing competition among domestic banks. The methods are inspired by Cornaggia, Mao, Tian, and Wolfed (2015). They employ the staggered deregulation of interstate bank branching laws that took eﬀect in 1994 in the US under the Interstate Banking and Branching Eﬃciency Act (IBBEA) as a shock to examine the impact of banking competition on innovation. They argue that banking competition enables small, innovative ﬁrms to secure ﬁnancing instead of being acquired by public corporations, which reduces the supply of innovative targets and makes the patents a trend of decline in the overall level. Unlike them, we use a DID approach to demonstrate the causality between foreign bank entry and enterprise innovation in China. We further illustrate that the consequence of foreign bank entry derives mainly from intensifying the competition among domestic banks, exerting its spillover eﬀect. We contribute to the existing literature in four ways. First, we provide detailed evidence on how ﬁnancial development aﬀects economic growth. Second, this paper focuses on the analysis and test of the spillover eﬀect of foreign bank entry, where By the end of 2013, a total of 51 countries and regions had set up 42 foreign-funded corporate bodies, 92 foreign bank branches and 187 representative oﬃces in China. Taking Bank of America for example, it took a stake in China Construction Bank for the ﬁrst time on 17 September 2007, holding an 8.52% stake, and the ratio rised up to 19.13% in 2008. 66 BAI ET AL. it promotes the allocation eﬃciency of domestic banks leading to the growth of innovation activities among domestic ﬁrms. This oﬀers a comprehensive understanding of the economic consequences of foreign bank entry and its mechanism. Third, we expand the research on enterprise innovation. Traditional literature mainly provides evidence at a ﬁrm level, industry level and market level. However, more and more scholars begin to realise that the institutional environment, especially the ﬁnancial environment which inﬂuences the enterprise innovation directly, should not be over- looked (Cornaggia et al., 2015; Hsu, Tian & Xu, 2014; Xie & Fang, 2011). Taking as an example a transitional economy like China, we try to explain how to reform the ﬁnancial system to serve and support the enterprise technology innovation better. Finally, the current literature on ﬁnance and economic growth suﬀers endogenous concern (to diﬀerent extents). To address this concern, we use the staggered multiple exogenous shocks and a DID approach to accurately identify the causality between them. The rest of the paper proceeds as follows. Section 2 describes the speciﬁc institutional background in China, develops theoretical analysis and puts forward a yet-to-be-tested hypothesis. Section 3 describes the data and variable construction. Section 4 presents the baseline results and endogeneity tests. Section 5 discusses the underlying mechan- isms of our baseline results, and Section 6 concludes. 2. Institutional background, theoretical analysis and hypotheses 2.1. Institutional background Since the economic reform and opening up policy was implemented in China in 1978, the reform of ﬁnancial markets has also seen signiﬁcant changes, such as the establish- ment of local banks and non-ﬁnancial institutions, division and going public for Big-4 state owned commercial banks. However, the banking market structure remains the same where the Big-4 state-owned banks dominate the bank market and privately- owned banks play a minor role in the Chinese banking market (Berger, Hasan, & Zhou, 2009). The institutional structure of the Chinese banking sector did not undergo an active and signiﬁcant change until China entered the WTO in 2001 (Alessandra & Poncet, 2008; Allen, Qian, Qian, & Zhao, 2009; Berger et al, 2009; Liu, 2008; Qian, Strahan, & Yang, 2015; Yao, Feng, & Jiang, 2011). According to the accession commitments to WTO at the end of 2001, the geographic and client restrictions on foreign bank lending in China were phased out gradually until 2006. With the deregulation, foreign banks set about business in China through various methods, such as setting up branches or subsidiaries and participating as strategic investors in Chinese banks (Figure 1). It constituted some staggered multiple exogenous shocks due to the gradual process of the banking sector’s opening-up, together with the fact that the timing and region choices of foreign bank entry were virtually decided by the central government rather than the needs of ﬁrms’ demand (Lin, 2011). Since then, foreign banks have become a new ﬁnancial source for Chinese ﬁrms. More importantly, China’s banking sector, which lagged far behind the product market, began to face unprecedented external pressure of competition. As a result, both the On 25 July 2014, three private banks- Webank, KinCheng Bank of Tianjin, and Civil and Commercial Bank of Wenzhou -were formally approved, meaning that China has a real sense of private banks. CHINA JOURNAL OF ACCOUNTING STUDIES 67 Figure 1. The process of foreign bank entry in China. government and banking sector took most aggressive measures. First, the government began to minimise the direct intervention of commercial banks’ operating in deposits and loans to make them economic entities of independent interests and operational autonomy, which enhanced their market competitiveness (Liu, 2008); second, the domi- nant large state-owned commercial banks’ joint-stock reform carried on until 2002. Some transnational banks and transnational corporations started to hold shares in China Construction Bank, Bank of China and Industrial and Commercial Bank of China. With the change of internal property structure, a new mechanism of supervision and management emerged, promoting the marketisation of the Chinese banking sector on a microeconomic foundation (Bai & Lian, 2012; Jian, Gan, & Yu, 2013). These reforms stimulated stronger proﬁt motivation and operational autonomy of the bank, promoted the competition in the banking sector and, subsequently, improved domestic banks’ economic behaviour and credit capital allocation. China’s marketisation reform of the banking sector began to enter a new era (Allen et a1., 2009; Jian et al., 2013; Yao et al., 2011), and the ﬁnancial environment faced by micro enterprises greatly improved. 2.2. Theoretical analysis and hypotheses As noted above, foreign bank entry has acted as an important event for China’s banking sector. Speciﬁcally, we sum up two paths through which foreign bank entry aﬀects domestic enterprise innovation (Figure 2): the direct eﬀect and the spillover eﬀect. the direct effect: increasing in the loan supply, high level of debt governance Foreign Domestic bank enterprise improve the efficiency of domestic banks entry innovation the spillover effect: intensifying the competition, improving domestic banks’ functions of financing and governance Figure 2. The mechanism of foreign bank entry aﬀecting domestic enterprise innovation. 68 BAI ET AL. The direct eﬀect of foreign bank entry is abundant capital and enhanced innovation ability, which renders the gathering and allocation of funds more eﬀective, improving the total credit in China (Zhou, 2012) and thus increasing the available funds for domestic enterprises. Using the standard credit supply model under the credit con- straints, Mao, Wu, and Liu (2010) ﬁnd that foreign bank entry has a signiﬁcant impact on corporate credit. Lin (2011) further demonstrates that foreign bank entry increases the long-term loans of high proﬁtability and non-state-owned enterprises, while it has no signiﬁcant inﬂuence on average. Furthermore, foreign bank entry goes hand-in-hand with the new technology, optimal allocation of resources and high banking system eﬃciency (Levine, 1996). Therefore, the credit of foreign banks also plays the role of debt’s corporate governance owing to its hard credit constraint. Compared with the direct eﬀect, the spillover eﬀect of foreign bank entry makes a lot of sense (Clarke, Cull, Peria, & Sánchez, 2005); more speciﬁcally, foreign bank entry improves the eﬃciency of domestic banks, which impacts on ﬁnancial behaviour and innovative activities. Ample literature shows that, as a result, foreign bank entry will ﬁght for deposits and clients, then alter the market structure of Chinese banking sectors and intensify competition. Owing to the strong intervention of the Chinese government to its original credit markets, the slow progress of marketisation reform, the government state of state-owned banks, the lack of eﬀective competition, foreign bank entry will show a higher marginal eﬀect on the promotion of banking competition. Wang and Song (2012) adopt structural disconnection to examine the eﬀect of China’s accession to the WTO on the Chinese banking sector. They ﬁnd that it signiﬁcantly intensiﬁes the domestic banking competition. Furthermore, according to the ‘performance hypothesis’, an increase in bank competition breaks the status of big state-owned banks’ governing markets, which make them apply the market rules better. For example, it forces them to change management decisions, upgrade technology and management level, and pro- vide better quality services (Degryse et al., 2009; Levine, 1996, 2004; Li & Han, 2008; Zhang & Wu, 2010). Due to the real economy relying on ﬁnancial development, and the credit ﬁnancing being the most important part of Chinese corporate ﬁnancing structure, when local banks’ eﬃciency is improved by foreign banks, plenty of micro enterprises will become the ultimate beneﬁciaries. In terms of the quantity, in order to protect market share, local big banks faced with competition from foreign banks will provide enterprises with more credit funds, especially those quality enterprises facing ﬁnancing constraints and having growth opportunities (Clarke et al., 2005). In regard to the quality, whether changes of initiative under foreign bank entry, or the subtle inﬂuence of foreign banks in the process of competition, the service quality and risk management level of domestic banks make substantial progress (Yao et al., 2011), thus local banks will better identify the high-quality business and high-quality project in the market. The important role that ﬁnancial development and structure features play on innovation activities both at the macro and the micro level is all the more high- lighted (Chava, Oettl, Subramanian, & Subramanian, 2013;Cornaggia et al., 2015;Hsu et al., 2014; Zhang, Yang, & Xin, 2016). However, due to ﬁnancial regulations in China, the development of the ﬁnancial market lags behind, which compels compa- nies to rely on indirect ﬁnancing channels to get the long-term and large-scale investment necessary for research and development (Zhang, Yang, & Xin, 2016). Given that the innovation activities of enterprises in China are undergoing a period CHINA JOURNAL OF ACCOUNTING STUDIES 69 of imitation and catching-up (Wu & Mi, 2011), the risk of technological and product innovation is still at a low level. It makes the banking sector compatible to the current stage of micro enterprise innovation activities (Lin, Sun, & Jiang, 2009;Zhang et al., 2016). Therefore, it is reasonable to expect that the institutional change in the banking sector may exert an important inﬂuence on the real-sector innovation activities in a transitional economy. Speciﬁcally, both the direct eﬀect of foreign bank entry and its spillover eﬀect on the eﬃciency of the local banking system are reﬂected in the optimisation of credit allocation and the reinforcement of debt governance. The additional credit supply from direct entering foreign banks allevi- ates the ﬁnancial constraints of businesses, especially those innovative ﬁrms with strong reliance on bank credits (Chava et al., 2013;Cornaggia et al., 2015). In the meantime, in the industrial sector represented by state-owned enterprises, the budget constraints become harder, which leads to the improvement of production eﬃciency. For example, the enterprises could adjust their production decisions to engage more in technological innovation activities (Bertrand, Schoar, & Thesmar, 2007). Jian et al. (2013) make use of the Chinese data and ﬁnd that the marketisation of the banking sector that began in 2001 greatly enhanced the total factor produc- tivity of micro enterprises, and that one of the important mechanisms is banking sector reform, promoting technological innovation of micro-enterprises. In conclusion, this paper proposes a yet-to-be-tested core hypothesis. That is, foreign bank entry signiﬁcantly promotes the innovation of domestic enterprises. The mechan- ism includes a direct eﬀect and a spillover eﬀect; the former is the foreign bank entry which directly increases the obtainable credit funds, while the latter is the foreign bank entry enhancing domestic banking competition, thus providing better service to the local enterprise innovation. 3. Sample, variables and model 3.1. Sample We obtain foreign bank entry data from the Regional Financial Operation Report of China. Due to data availability, the paper restricts the sample to publicly-traded ﬁrms. For the enterprise innovation, we obtain data from the China’s Listed Companies Patent Research database, which is based on The State Intellectual Property Oﬃce of the People’s Republic of China and provides patent applications and the patent situation of all publicly-traded ﬁrms in the Chinese Stock Market. Additionally, for the sample ﬁrms with available enterprise innovation data, we obtain ﬁnancial data from the China Stock Market and Accounting Research database (CSMAR), scientiﬁc researchers’ data from the RESSET database and controlling shareholder data from the Center of China Economic Research Service database (CCER). The banking sector’s restrictions in China were phased out gradually from 2001 to 2006 and our analysis covers the period of 2001–2014. At the same time, considering that the technology innovation activities are mainly distributed in the manufacturing and high-tech industry, we choose manufactur- ing and the information technology industry as the initial samples. Our ﬁnal samples are The annual Regional Financial Operation Report of China is published by the People’s Bank of China (PBOC). The report contains banking data of each province in China and can be obtained from the oﬃcial website of PBOC. 70 BAI ET AL. unbalanced panel data consisting of 9993 ﬁrm-year observations after excluding the ﬁrms of ST, *ST and observations with missing values. Table 1 shows the distribution of industry as time changes. 3.2. Variables 3.2.1. Enterprise innovation In general, enterprise innovation is captured from the input or output aspect. The former includes the innovation elements’ inputs, such as R&D, and scientiﬁc and technical personnel, while the latter includes the number of patents, and new product sales income. However, the indicator of innovation input cannot fully reﬂect the actual innovation level of the enterprise, since it only reﬂects the input in the process of innovation (Wu, 2007). In addition, with the major data of R&D missing before 2007, more and more papers have begun to use innovation output for related research (Chava et al., 2013; Cornaggia et al., 2015; Li & Zheng, 2016; Tan, Tian, Zhang, & Zhao, 2014; Tong, He, He, & Lu, 2014; Wen & Feng, 2012; Zhou, Cheng, & Wang, 2012). Accordingly, we use the number of patent applications ﬁled in a year to reﬂect the innovation at the ﬁrm level. Based on the information available in China’s Listed Companies Patent Research Database, we construct enterprise innovation at the ﬁrm-year level through two steps. (1) To reﬂect the long-term nature of investment in innovation, following Cornaggia et al. (2015), we consider the ﬁrm’s total patents generated in the next three years. This approach mitigates the inﬂuence of idiosyncratic shocks that could distort the innovation output in any particular year. (2) The distribution of patents in the sample is right skewed. Therefore, we use the natural logarithm of the total number of patents generated in the subsequent three years, LnPat, as the main enterprise innovation Table 1. Distribution of sample’s industry and year. 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Total C0 48 50 51 49 50 51 58 63 67 78 84 649 C1 45 46 48 54 55 57 59 61 64 72 79 640 C3 21 20 19 23 22 24 27 31 38 42 47 314 C4 123 125 125 135 129 133 137 156 162 195 226 1646 C5 30 32 37 41 42 50 61 67 73 105 130 668 C6 91 100 104 109 111 112 117 121 127 157 186 1335 C7 155 160 162 177 177 185 204 225 264 351 417 2477 C8 70 71 72 84 88 89 82 85 98 118 129 986 C9 15 16 16 19 19 21 24 26 28 34 39 257 G 61 66 72 73 75 72 75 82 111 152 182 1021 Total 659 686 706 764 768 794 844 917 1032 1304 1519 9993 Note: According to the industry classiﬁcation guideline of the China Securities Regulatory Commission (CSRC) and Fang (2009), we classify manufacturing industry at two codes and other industries at one code. Speciﬁcally, C0 is food and beverage, C1 is textile, garments and fur, C2 is wood, furniture and other manufacturing, C3 is papermaking and printing, C4 is petroleum, chemical, plastic and plastic, C5 is electronic, C6 is metal and non-metal, C7 is machinery, equipment and instrument, C8 is medicine and biological products, C9 is other manufacturing, G is information technology. The sample interval is 2001–2014 as the dependent variable is related to the aggregation of three years’ data. ST means that the company has suﬀered losses for two consecutive years, and *ST means that the company has suﬀered losses for three consecutive years. We use a patent’s application year instead of its grant year because the application year better captures the actual time of innovation (Cornaggia et al., 2015). CHINA JOURNAL OF ACCOUNTING STUDIES 71 measure in our analysis. In particular, to avoid losing observations with zero patents, we add one to the actual values when calculating the natural logarithm. Furthermore, according to the Patent Law of the People’s Republic of China (2008), we classify patents into invention, utility model and appearance design. We regard invention as being truly innovation while the utility model and appearance design are usually revisions or adaptations of an existing patent, also called a ‘gizmo’ (Li & Zheng, 2016). Following Chen et al. (2015), we construct a second measure of enterprise innovation (LnPat_cut) by dividing patents into invention, utility model and appearance design, and weighting a discount of 50%, 30% and 20% respectively. The LnPat captures the quantity of innovation output while the LnPat_cut captures the quality of innovation output. In order to avoid subjective interference in the process of weighting coeﬃcients, we follow Li and Zheng (2016) and employ these three types of patent as dependent variables directly instead of LnPat and LnPat_cut when conducting robustness tests. 3.2.2. Foreign bank entry To identify the impact of foreign bank entry on domestic enterprise innovation, the paper exploits the time series variation induced by the change in the ﬁnancing environ- ment because of foreign bank entry and the cross-sectional variation induced by the geographic location of domestic ﬁrms, based on the WTO banking sector liberalisation schedule. We construct two measures for a province-year’s foreign bank entry. The ﬁrst measure, Enter_p, represents a dummy variable that is equal to one when foreign banks are available in the region and zero otherwise. Given the fact that foreign banks may not run renminbi business even after their entry, we construct another measure of foreign bank entry, Enter_r, which is equal to one if foreign banks engage in renminbi business in the province, and zero otherwise. In the robustness test, following Yao et al. (2015), we also use the proportion of foreign banks’ branches, workers and assets accounted for those of each area to measure foreign bank entry. 3.2.3. Other control variables Following Cornaggia et al. (2015), Chava et al. (2013), Li and Zheng (2016), Tan et al. (2014), Tong et al. (2014) and Wu (2007), we include control variables to capture possible confounding eﬀects as follows: the ratio of net value of intangible asset to total assets (Intangible), the natural logarithm of a ﬁrm’s number of scientiﬁc researchers (Staﬀ), the natural logarithm of the year-end values of a ﬁrm’s total assets (Size), the number of years the ﬁrm has existed since the founding year (Age), the ratio of net value of cash ﬂow from operating activities to total assets (Cfo), the ratio of net value of ﬁxed assets to total assets (Tangible), the ratio of net income to total assets (Roa), the book value to total assets (TobinQ), whether it is a State-owned enterprise (SOE) or not (Soe), and the Herﬁndahl-Hirschman Index of Industry (HHI). Detailed deﬁnitions of the variables are shown in Table 2. To minimise the eﬀect of outliers, we winsorise all variables at the ﬁrm level that have the potential to be unbounded at the 1st and 99th percentiles of their empirical distributions. The reasons we use the variable of foreign bank entry as the main test instead of its degree are as follows: ﬁrst, the former is more exogenous while the latter is susceptible by the development level of the regional economy. Second, once foreign banks enter, they will have a signiﬁcant eﬀect on the market expectation of state-owned banks and facilitate competition. In this case, the size of corporates is not important anymore. 72 BAI ET AL. Table 2. Deﬁnition of variables. Variables Deﬁnition LnPat Enterprise innovation. It equals the natural logarithm of one plus ﬁrm’s total number of patents ﬁled in years t+1 through t+3. LnPat_cut Enterprise innovation. It is measured as the natural logarithm of one plus the ﬁrm’s total number of inventions, utility models and appearance designs in years t+1 through t+3 separately at a discount of 50%, 30% and 20%. Enter_p An indicator variable that takes the value of one if foreign banks are permitted to enter and zero otherwise Enter_r An indicator variable that takes the value of one if foreign banks are allowed to do local-currency business loans with domestic ﬁrms and zero otherwise. Intangible Intangible asset is measured as the ratio of net value of Intangible asset to total assets. Staﬀ Firm staﬀ is measured as the natural logarithm of a ﬁrm’s number of scientiﬁc researchers. Size Firm size is measured as the natural logarithm of a ﬁrm’s total assets of its closing values. Age Firm’s age. It equals the number of years the ﬁrm has existed since the founding year. Cfo Cash ﬂow is measured as the ratio of net value of cash ﬂow from operating activities to total assets. Tangible Tangible asset is measured as the ratio of the net value of ﬁxed assets to total assets. Roa Return on assets is measured as the ratio of net income to total assets. TobinQ Tobin’s Q is measured as the sum of market value of shares and liabilities divided by book value to total assets, substituting net assets per share for market value of non-tradable shares. Soe A dummy variable equals one if the ﬁrm is a SOE and zero if the ﬁrm is a non-SOE. HHI Herﬁndahl-Hirschman Index, HHI=∑ (X /∑ X ) , X is sales volume of a ﬁrm f. f f f f Year A dummy variable equals one if the ﬁrm went public during that year, and zero otherwise. 3.3. Summary statistics Table 3 reports the summary statistics of the ﬁnal samples used to test the hypotheses. Panel A provides descriptive statistics in the ﬁrm-year level. On average, the natural logarithm of patents peer ﬁrm in the subsequent three years is 2.0774, and it decreases to 1.5101 after considering the discount of utility models and appearance designs. The average value of Enter_p (Enter_r) is 81.49% (71.33%). Panel B presents the comparisons descriptive statistics for the sample. The value of the natural logarithm of patents peer ﬁrm in the subsequent three years in the province where there was no foreign bank entry is 1.1212 (1.2889) on average, while it is 2.2947 (2.3943) in the province where foreign banks enter. The results remain robust when considering the discount of utility models and designs. Thus, based on the univariate comparison, we conclude that the ﬁrms located in the place where foreign banks enter are more likely to engage in innovation. 3.4. Baseline speciﬁcation As we discussed in the introduction, a major challenge of our study is the identiﬁcation of the causal eﬀects of foreign bank entry on domestic enterprise innovation, due both to omitted variables and to reverse causality concerns. First, unobservable province characteristics related to both foreign bank entry and enterprise innovation could remain in the residual term of regressions. These unobservable characteristics make it diﬃcult to draw unbiased statistical inferences from standard OLS regressions. Second, there is an existing debate on the direction of causality between ﬁnance and economic growth (see Butler & Cornaggia, 2011, for a review). Fortunately, the staggered entry of foreign banks in China generates plausibly exogenous variation in a province’s ﬁnancial environment (Lin, 2011), which allows us to overcome a common diﬃculty. Speciﬁcally, we estimate the following model: CHINA JOURNAL OF ACCOUNTING STUDIES 73 Table 3. Summary statistics. Variables N Mean Std. Min Q1 Median Q3 Max Panel A：Descriptive statistics LnPat 9993 2.0774 1.9069 0.0000 0.0000 2.1972 3.5553 6.8835 LnPat_cut 9993 1.5101 1.5053 0.0000 0.0000 1.3350 2.6027 5.8409 Enter_p 9993 0.8149 0.3884 0.0000 1.0000 1.0000 1.0000 1.0000 Enter_r 9993 0.7133 0.4522 0.0000 0.0000 1.0000 1.0000 1.0000 Intangible 9993 0.0373 0.0395 0.0000 0.0106 0.0262 0.0501 0.2175 Staﬀ 9993 5.5743 1.1787 2.7081 4.7958 5.5334 6.2953 8.7910 Size 9993 21.3392 1.0249 19.3296 20.6288 21.1949 21.9023 24.6525 Age 9993 9.9912 4.5057 2.0000 7.0000 10.0000 13.0000 22.0000 Cfo 9993 0.0458 0.0739 -0.1700 0.0062 0.0445 0.0875 0.2614 Tangible 9993 0.2679 0.1576 0.0098 0.1447 0.2420 0.3705 0.6771 Roa 9993 0.0373 0.0605 -0.2243 0.0142 0.0383 0.0654 0.1967 TobinQ 9993 1.6913 0.8969 0.5301 1.1597 1.3993 1.8690 6.1222 Soe 9993 0.5803 0.4935 0.0000 0.0000 1.0000 1.0000 1.0000 HHI 9993 0.0531 0.0388 0.0176 0.0311 0.0385 0.0575 0.1796 Panel B：Mean comparison Variables Enter_p=0 Enter_p=1 Diﬀ Enter_r=0 Enter_r=1 Diﬀ LnPat 1.1212 2.2947 –1.1734*** 1.2889 2.3943 –1.1054*** LnPat_cut 0.7354 1.6862 –0.9507*** 0.8723 1.7665 –0.8943*** Note: The test for mean diﬀerence is t test. *, **, and ***indicate the 0.1, 0.05, and 0.01 levels of signiﬁcance, respectively, for a two-tailed test. Patent ¼ α þ α Enter þ α EFD þ α Enter EFD i;tþ1 toþ3 0 1 j;t 2 k 3 j;t k þ Control þ μ þ Year þ ε (1) i;t t i;t where i indexes ﬁrm, t indexes time, j indexes province, and k indexes industry. The dependent variable, Patent, includes two indicators in model (1): LnPat and Lnpat_cut.Here LnPat is deﬁned as the natural logarithm of the number of the ﬁrm’s patents in the following three years. Lnpat_cut is deﬁned as the natural logarithm of the number of the ﬁrm’s patents in the following three years after considering the discount of utility models and appearance designs. Enter is the foreignbankentry,weuse twovariables to measureitineachprovinceineach year: Enter_p and Enter_r. EFD denotes the degree of reliance on external capital of the sector prior to foreign bank entry. Industries that were more dependent on bank ﬁnancing prior to the reform were more exposed to the distorted lending practices, and therefore should be more sensitive to foreign bank entry. Accordingly, we use EFD to distinguish the treat group and control group, and measure it with the mean value of all the ﬁrms in the same industry, following Bertrand et al. (2007)and Jian et al.(2013). The coefﬁcient of Enter*EFD (Enter_p*EFD or Enter_r*EFD)is α , which captures the net eﬀect of foreign bank entry on domestic enterprise innovation. The Control comprises a set of control variables that could aﬀect enterprise innovation, including Intangible, Staﬀ,Size, Age, Cfo, Tangible, Roa, TobinQ, Soe and HHI.Weinclude ﬁrm ﬁxed eﬀects, μ,and year eﬀects, Year, in the baseline regression to address the concern that In fact, the enterprises that have less degree of dependence on external capital face the same inﬂuence brought by foreign bank entry. It causes underestimation of the coeﬃcients but has no inﬂuence on the conclusions of this paper. Further discussion about the rationality and deﬁciencies of grouping, refer to Jian et al. (2013). 74 BAI ET AL. unobservable variables omitted from model (1) might be correlated with innova- tiveness, rendering our ﬁndings spurious. 4. Empirical results 4.1. Baseline results Table 4 reports the estimation results based on model (1). In column (1) and column (2), the coeﬃcient estimates of Enter*EFD (Enter_p*EFD and Enter_r*EFD) are positive and signiﬁcant at a 1% level. This ﬁnding suggests that foreign bank entry leads to an increase in the number of domestic ﬁrm’s patents in the ﬁrst three subsequent years, excluding the interference of other factors. Next, we examine the eﬀect of foreign bank entry on patent quality by replacing the dependent variable with LnPat_cut, the natural logarithm of domestic ﬁrm’s patents after considering the discount of utility models and appearance designs. We continue to observe positive and signiﬁcant coeﬃcient esti- mates of Enter*EFD at a 1% level in columns (3) and (4), suggesting that foreign bank Table 4. Foreign bank entry and domestic enterprise innovation: baseline results. Dependent variable LnPat LnPat_cut Enter_p Enter_r Enter_p Enter_r Enter= (1) (2) (3) (4) Enter –0.6889*** –0.6690*** –0.5998*** –0.5753*** (0.1538) (0.1549) (0.1205) (0.1214) EFD –2.5595 –2.3699 –1.7412 –1.6125 (2.9500) (2.9511) (2.3119) (2.3120) Enter*EFD 1.7760*** 1.6705*** 1.4810*** 1.4136*** (0.3130) (0.3184) (0.2453) (0.2494) Intangible 0.7890** 0.7864** 0.6307** 0.6295** (0.3193) (0.3195) (0.2502) (0.2503) Staﬀ 0.0866*** 0.0851*** 0.0805*** 0.0792*** (0.0146) (0.0146) (0.0114) (0.0114) Size 0.1796*** 0.1840*** 0.1692*** 0.1721*** (0.0246) (0.0246) (0.0193) (0.0193) Age 0.1290*** 0.1336*** 0.0949*** 0.0975*** (0.0405) (0.0405) (0.0317) (0.0317) Cfo –0.0989 –0.1047 –0.0831 –0.0883 (0.1408) (0.1409) (0.1104) (0.1104) Tangible 0.4369*** 0.4617*** 0.3918*** 0.4097*** (0.1059) (0.1059) (0.0830) (0.0830) Roa 0.3441* 0.3602* 0.2747* 0.2898** (0.1871) (0.1873) (0.1466) (0.1467) TobinQ 0.0168 0.0175 0.0178 0.0182 (0.0145) (0.0146) (0.0114) (0.0114) Soe 0.2527*** 0.2502*** 0.2362*** 0.2351*** (0.0460) (0.0460) (0.0361) (0.0361) HHI 0.4840 0.4754 1.2276** 1.2177** (0.6190) (0.6194) (0.4851) (0.4853) Constant –2.5461* –2.7321* –3.0025** –3.1270*** (1.5153) (1.5146) (1.1875) (1.1866) Firm & Year F.E. Yes Yes Yes Yes N 9993 9993 9993 9993 Within-R 0.2057 0.2047 0.2464 0.2459 F 93.3013 92.7190 117.7345 117.4486 Note: We cluster standard errors by ﬁrm in our all tests. *, **, and ***indicate the 0.1, 0.05, and 0.01 levels of signiﬁcance, respectively, for a two-tailed test. CHINA JOURNAL OF ACCOUNTING STUDIES 75 entry results in an increase in the quality of patents generated by the companies headquartered in the province in the subsequent three years. We also apply the variable of foreign bank entry’s duration to capture its time eﬀect on enterprise innovation. Speciﬁcally, we calculate Duration, the periods between the reporting year and the year of foreign bank entry for each province (Enter_p or Enter_r), and then get the interaction term of Duration and Enter*EFD (Enter_p*EFD or Enter_r*EFD). As is shown in Table 5, the coeﬃcients of Enter*EFD*Duration are signiﬁ- cantly positive, which indicates that the eﬀect of foreign bank entry on local enterprise innovation is increasing with longer duration and more outlets. 4.2. Robustness tests As discussed above, although the staggered entry of foreign banks in China represents an exogenous shock to the domestic banking system, we are still concerned that our basic results could be biased due to both omitted variables and reverse causality. Furthermore, foreign bank entry may be not strictly exogenous. To further explore those possibilities, we conducted robustness tests as follows. 4.2.1. Omitted variables One concern that prevents us drawing a causal interpretation of foreign bank entry on enterprise innovation from our baseline regressions is an omitted variables problem: unobservable shocks or variables that are omitted from our analysis but coincide with foreign bank entry could drive our results. To address this concern, we conduct a placebo test to check whether our results are aﬀected when we artiﬁcially (i.e. incor- rectly) assume the foreign bank entry in years other than the actual year of foreign bank entry. We do this by assigning a province randomly into each of these entry years. This approach disrupts the proper assignment of foreign bank entry years to province. Table 5. Foreign bank entry and domestic enterprise innovation: the time eﬀect. Dependent variable LnPat LnPat_cut Enter_p Enter_r Enter_p Enter_r Enter= (1) (2) (3) (4) Enter*EFD 0.3611*** 0.2247*** 0.2953*** 0.2000*** (0.0900) (0.0691) (0.0705) (0.0542) Duration –0.1352*** –0.1357*** –0.0953*** –0.0943*** (0.0292) (0.0289) (0.0229) (0.0227) Enter*EFD*Duration 0.2876*** 0.2388*** 0.2274*** 0.1753*** (0.0501) (0.0584) (0.0393) (0.0458) Constant –3.7220*** –3.8122*** –3.8204*** –3.8929*** (0.5431) (0.5432) (0.4257) (0.4258) Control variables Yes Yes Yes Yes Firm & Year F.E. Yes Yes Yes Yes N 9993 9993 9993 9993 Within-R 0.2070 0.2052 0.2473 0.2455 F 93.9992 93.0098 118.3362 117.1657 Note: We cluster standard errors by ﬁrm in our all tests. *, **, and ***indicate the 0.1, 0.05, and 0.01 levels of signiﬁcance, respectively, for a two-tailed test. 76 BAI ET AL. Therefore, if an unobservable shock (or province characteristics) occurs at approximately the same time as the foreign bank entry during 2001–2006, it should still reside in the testing framework, and thus have an opportunity to drive the results. However, if no such shock exists, then our incorrect assignments of entry years of foreign banks to provinces should weaken our results when we re-estimate our baseline regressions in model (1). Table 6 shows the results of the placebo test. The coeﬃcient estimates of Enter*EFD is statistically insigniﬁcant. Therefore, we rule out the possible inﬂuence of omitted variables. 4.2.2. Reverse causality Zhang and Yang (2007) argue that market opportunity is one of the important factors aﬀecting the location choice of foreign banks. Consequently, another concern, reverse causality, may arise if provinces diﬀer in their innovation intensities and such diﬀerences triggered the foreign bank entry. To explore the possibility of reverse causality, we examine the dynamics of innovation surrounding foreign bank entry, following Chava et al. (2013) and Cornaggia et al. (2015). If reserve causality is indeed present, we should observe changes in innovation prior to foreign bank entry. We decompose each province associated with four periods around the foreign bank entry year: all years up to and including two years prior to foreign bank entry, one year prior to foreign bank entry, one year post foreign bank entry, and two years or more 2+ 1 1 post foreign bank entry. We then obtain four new variables, Before , Before , After , and 2+ After , corresponding to the four time periods around each foreign bank entry. The year of foreign bank entry (Enter_p or Enter_r) is the reference year in this setting. The 2+ 1 coeﬃcient estimates of Before and Before are especially important because their signiﬁcance and magnitude would indicate whether domestic enterprise innovation causes foreign bank entry. We estimate the following model: Table 6. Foreign bank entry and domestic enterprise innovation: placebo test. Dependent variable LnPat LnPat_cut Enter_p Enter_r Enter_p Enter_r Enter= (1) (2) (3) (4) Enter –0.0533 0.0172 0.0178 0.0191 (0.1324) (0.1178) (0.1037) (0.0923) EFD –1.0301 –0.9337 –0.3953 –0.3989 (2.9536) (2.9491) (2.3146) (2.3110) Enter*EFD 0.0897 –0.0581 –0.0484 –0.0617 (0.2739) (0.2440) (0.2146) (0.1912) Constant –3.6654** –3.7175** –3.9547*** –3.9509*** (1.5131) (1.5104) (1.1857) (1.1836) Control variables Yes Yes Yes Yes Firm & Year F.E. Yes Yes Yes Yes N 9993 9993 9993 9993 Within-R 0.2005 0.2005 0.2415 0.2415 F 90.3213 90.3236 114.6685 114.6964 Note: We cluster standard errors by ﬁrm in our all tests. *, **, and ***indicate the 0.1, 0.05, and 0.01 levels of signiﬁcance, respectively, for a two-tailed test. Although the dependent variable we use is at ﬁrm level, considering the degree of enterprise innovation highly related to the total innovations of the region it belongs, our concerns are necessary. CHINA JOURNAL OF ACCOUNTING STUDIES 77 Table 7. Foreign bank entry and domestic enterprise innovation: reverse causality. Dependent variable LnPat LnPat_cut Enter_p Enter_r Enter_p Enter_r Enter= (1) (2) (3) (4) 2+ Before 0.0278 -0.0302 0.0398 -0.0160 (0.0480) (0.0425) (0.0438) (0.0388) Before 0.0125 -0.0121 0.0129 -0.0119 (0.0518) (0.0489) (0.0473) (0.0445) After 0.0552 0.0480 0.0433 0.0415 (0.0510) (0.0474) (0.0465) (0.0432) 2+ After 0.1540*** 0.1239*** 0.1303*** 0.1125*** (0.0431) (0.0395) (0.0393) (0.0360) Constant 0.7040*** 0.7407*** 0.5822*** 0.6174*** (0.0477) (0.0456) (0.0435) (0.0416) Firm & Year F.E. Yes Yes Yes Yes N 15132 15132 15132 15132 Within-R 0.1418 0.1420 0.1575 0.1576 F 128.48 128.61 143.52 143.74 Note: We cluster standard errors by ﬁrm in our all tests. *, **, and ***indicate the 0.1, 0.05 and 0.01 levels of signiﬁcance, respectively. Since we changed the measuring method of variables, the sample diﬀer with the previous ones. 2þ 1 1 2þ Patent ¼ β þ β Before þ β Before þ β After þ β After i;t i;t i;t i;t i;t 0 1 2 3 4 (2) þμ þ Year þ ε t i;t Table 7 reports the estimation results based on model (2). In column (1) and column (2), we report the results for enterprise innovation measured by the natural logarithm of 2+ 1 patents. The coeﬃcient estimates of Before and Before are not signiﬁcant, suggesting that enterprise innovation shows no signiﬁcant change prior to foreign bank entry. The 1 2+ coeﬃcient estimate of After is positive but not signiﬁcant while After is positive and signiﬁcant. (This result is reasonable because the impact of foreign bank entry on domestic enterprises is lagged, especially for enterprises’ activities with a long cycle. The basic model of this article uses the ﬁrm’s total patents generated in the next three years as the dependent variable for the same reason. Cornaggia et al., 2015, also found that the change of enterprise innovation is very weak after the ﬁrst year of banking deregulation.) In columns (3) and (4), we repeat these tests with the natural logarithm of patents after considering the discount of utility models and appearance designs and we observe a similar pattern around the year of foreign bank entry. Overall, the results presented in Table 7 suggest that, whether we measure enterprise innovation with LnPat or LnPat_cut, there is not a pre-existing trend in innovation before the year of foreign bank entry. These results mitigate concerns about reverse causality. 4.2.3. Additional robustness checks To make our results more convincing, a rich set of other tests are carried out in this section. (1) Based on the basic model, the time-variation trend of each province, and the interaction items of the province dummy variable and the year dummy variable are employed to control the possible temporal trend diﬀerence. (2) A more exogenous shock, i.e. the ﬁnancial sector’s opening to the outside comprehensively in 2006, is used to construct another DID model in this paper. We stress that before 2006 there were provinces where some foreign banks entered while others did not. Since 2006, foreign banks have freely had access to any province in China. Owing to the fact that the DID approach needs to be based on a parallel trend 78 BAI ET AL. between the treatment group and the control group, we construct the model by observing the net eﬀect due to a lack of foreign banks before 2006 (rather than entering). We expect to see the opposite result when foreign banks have not entered yet. (3) Although the ﬁrm’s individual ﬁxed eﬀects are controlled, the innovation outputs will inevitably be aﬀected by other factors, such as macro-level factors (regional economic development level and ﬁnancial development level), ﬁrm-level factors (corporate capital structure, ownership structure, and the size of the board and its structure, executive incentives). With the aforementioned variables controlled, we regress again and get similar results. (4) We employ the panel ﬁxed eﬀects model, excluding interaction item, to regress foreign bank entry and enterprise innovation directly (Cornaggia et al., 2015). (5) We classify the patent into three types: invention, utility model and appearance design, and ﬁnd that foreign bank entry facilitates the ﬁrsttwo kindsofpatentbuthasnosigniﬁcant change on the number of appearance designs when we employ these three types as dependent variables, which is consistent with previous ﬁndings. (6) We follow Yao et al. (2015) to measure the degree of foreign bank entry by using its ratio of branches, employees and assets accounted for those of each area. All the above is consistent with previous ﬁndings. 5. Mechanism analysis: the direct eﬀect and the spillover eﬀect of foreign bank entry First, in terms of the direct eﬀect of foreign bank entry, not only does it bring in enough capital, but excellent business innovation ability can also be more eﬀective in the collection and distribution of funds, which will increase the availability of credit funds from local enterprises and support them in innovation activities. To test this eﬀect, we use local enterprises to obtain loans from foreign banks or not (Loan_dum) and the ratio of foreign bank loans to total assets (Loan) as the explanatory variable for regression analysis. From Table 8 Panel A, we ﬁnd that the coeﬃcient of Enter_r*EFD is not signiﬁcantly positive and the coeﬃcient of Enter_r*EFD is signiﬁcantly positive when Loan_dum is taken as an explanatory variable. In addition, the coeﬃcients are similar when Loan is taken as an explanatory variable. Overall, after foreign bank entry (espe- cially when foreign banks carry out renminbi business), the possibility and the scale of obtaining loans from foreign banks has increased to enterprises that excessively rely on external ﬁnancing. Furthermore, Table 8 Panel B reports the possibility of an enterprise obtaining foreign bank loans and the regression result of the access scale of foreign banks’ loans and enterprise innovation. The results show that the coeﬃcients are positive, but only column (1) and column (2) are signiﬁcant at the 10% level. It illustrates that foreign bank entry increases the possibility and scale of obtaining loans from foreign banks, but the promoting eﬀect to innovation remains limited. Second, as previously analysed, the spillover eﬀect of foreign bank entry is more prominent. We champion the idea that the nurturing role of foreign bank entry on enterprise innovation derives mainly from its spillover eﬀect by enhancing competition among domestic banks. The data come from the database of bank loans of CSMAR listed companies in China, including the issuing bank, the amount, the period, and the interest rate. Thus, it can identify whether the listed companies obtain loans from foreign banks or not and the exact amount. What needs to be explained is that most of the listed companies did not disclose the speciﬁc name of the issuing bank. For the sake of rigour, only samples with accurate bank identiﬁcation are retained and, as a result, the sample size has declined. CHINA JOURNAL OF ACCOUNTING STUDIES 79 Table 8. Foreign bank entry and domestic enterprise innovation: the direct eﬀect. Dependent variable Loan_dum Loan Enter_p Enter_r Enter_p Enter_r Enter= (1) (2) (3) (4) Panel A Enter*EFD 32.6948 4.6229** 8.0553 1.5090** (2522.919) (2.2013) (402.3202) (0.6921) Constant –46.4700 –34.2471 –15.5242 –12.8911 (1607.59) (499.5682) (280.2898) (162.1547) Control variables Yes Yes Yes Yes Industry & Year F.E. Yes Yes Yes Yes N 9307 9307 9307 9307 Pseudo R 0.1817 0.1898 0.1821 0.1893 LR chi 140.77 147.06 141.12 146.68 Panel B Loan_dum –2.8092** –1.8631* (1.3896) (1.0092) Enter*EFD*Loan_dum 5.9436* 3.9881* (3.1089) (2.3350) Loan –60.6903 –59.0130 (47.7540) (41.3294) Enter*EFD*Loan 132.7492 129.8708 (108.2396) (95.4869) Constant –3.5485** –3.5985** –3.6350** –3.6437** (1.4319) (1.4370) (1.4274) (1.4286) Control variables Yes Yes Yes Yes Industry & Year F.E. Yes Yes Yes Yes N 9993 9993 9993 9993 Within-R 0.2062 0.2050 0.2045 0.2046 Note: (1) The dependent variables in ﬁrst two column of Panel A are dummy variables, so we use logit regression, and the dependent variables of column (3) and column (4) have a large number of zeros, so we use probit regression; (3) The numbers in brackets are robust standard errors, *, **, and ***indicate the 0.1, 0.05, and 0.01 levels of signiﬁcance, respectively. Our empirical process includes two steps. First, we measure banking competition, Comp,by employing the ‘ﬁnance market competition index’. Thelargerthe index, the moreintense will be the competition of banks. Second, we establish a regression between the bank competition and foreign bank entry to measure the change of bank competition index resulting from foreign bank entry (Comp_e_p or Comp_e_r). Then we regress enterprise innovation (LnPat or LnPat_cut) on the change in bank competition caused by foreign bank entry (Comp_e_p or Comp_e_r). Table 9 reports the regression results. In columns (1) and (2), the coeﬃcient estimates of Enter are all positive and signiﬁcant at the 1% level, suggesting that foreign bank entry does enhance the degree of competition in the domestic banking sector (Comp). In columns (3), (4), (5) and (6), the coeﬃcient estimates of Comp_e are all positive and signiﬁcant at the 1% level, indicating that due to the boost of competition among domestic banks brought by foreign bank entry, the level of domestic enterprise innovation indeed increases and the conclusion is not aﬀected for other measures of innovation. In addition, we replace the banking competition with a new proxy variable, which is measured as the average assets of each banks’ branch. Conversely, the smaller the index, the more intense will be the competition The index from China’s marketization index of regional market relative processes reported in 2011 complied by Fan, Wang, and Zhu (2011), which is measured as the proportion of the non-state ﬁnancial institutions’ deposits accounted for all ﬁnancial institutions’ deposits. 80 BAI ET AL. Table 9. Foreign bank entry and domestic enterprise innovation: the spillover eﬀect. First step Second step Dependent Dependent variable Comp variable LnPat LnPat_cut Enter= Enter_p Enter_r Compe_e= Comp_e_p Comp_e_r Comp_e_p Comp_e_r (1) (2) (3) (4) (5) (6) Enter 0.2093*** 0.3570*** Comp_e 0.7755*** 0.3566*** 0.6687*** 0.3367*** (0.0276) (0.0256) (0.1622) (0.0869) (0.1500) (0.0803) Constant –14.2100*** –17.8577*** Constant –4.0396*** –4.0616*** –4.1236*** –4.1335*** (2.1778) (2.1827) (0.5423) (0.5425) (0.5015) (0.5015) Control Yes Yes Control Yes Yes Yes Yes variables variables Firm & Year F.E. Yes Yes Firm & Year F.E. Yes Yes Yes Yes N 9993 9993 N 9993 9993 9993 9993 2 2 Within-R 0.7374 0.7416 Within-R 0.0385 0.0378 0.0707 0.0704 F 1163.33 1189.17 F 100.28 99.92 117.43 117.29 Note: (1) According to prior research, we control a series of variables in the ﬁrst stage of regression such as Total industrial output of the area, Fixed assets investment, Proportion of primary industry, Proportion of secondary industry, Scale of imports and exports, Education level, Marketing process and Non-state economy development; (2) We cluster standard errors by ﬁrm in our all tests. *, **, and ***indicate the 0.1, 0.05 and 0.01 levels of signiﬁcance, respectively. of banks. The results are similar and we don’t report them. Overall, the regressions support our view that the nurturing role of foreign bank entry on domestic enter- prise innovation derives mainly from spillover eﬀects by enhancing competition among domestic banks. 6. Conclusions This article studies the eﬀects of foreign bank entry on domestic enterprise innovation and its mechanism of action. Starting with reviewing the process of foreign bank entry after China’s accession to WTO, we theoretically identify the mechanism of how foreign bank entry aﬀects China’s enterprise innovation, both from the direct and the spillover eﬀect. Then, we examine the relationship between foreign bank entry and domestic enterprise innovation, by exploiting the staggered multiple exogenous shocks and by using a DID approach and the data of Chinese listed companies. The results show that foreign bank entry signiﬁcantly promotes the innovation of domestic enterprises. Further study illustrates that the nurturing role derives not only from the direct eﬀect (although it is very limited), but also from the spillover eﬀect of foreign bank entry by enhancing competition among domestic banks. Foreign bank entry not only provides access to capital, but also pushes a series of reforms among domestic banks to accelerate China’s banking sector marketisation, eﬀectively driving the development of the real economy. Therefore, it is necessary to, open the door to ﬁnancial sector wider and lower the barriers of foreign bank entry. Our conclusions also provide some beneﬁcial enlightenment for the growing development of Chinese private banks. The paper emphasises the facilitation eﬀect of foreign bank entry on enterprise innovation, without involving the possible impacts on the Chinese ﬁnancial system’s stability, or monetary policy transmission. Therefore, we still need to understand the role of foreign bank entry in an objective and overall manner. CHINA JOURNAL OF ACCOUNTING STUDIES 81 Acknowledgements The authors thank the constructive suggestions from participants of the CJAS academic seminar held in 2016. The authors would like to oﬀer their most sincere thanks to the editors and the reviewers for their insightful comments. In addition, the authors specially thanks Dr Liang Han, at the University of Reading, UK, for his valuable advice and proofreading. The authors are respon- sible for any remaining errors. Funding This work was supported by the National Natural Science Foundation of China [71762026, 71562029];Fund of the Key Research Center of Humanities and Social Sciences in the General Colleges and Universities of Xin Jiang Uygur Autonomous Region [XJEDU020115B02]; References Alessandra, G., & Poncet, S. (2008). Could ﬁnancial distortions be no impediment to economic growth after all? Evidence from China. Journal of Comparative Economics, 36, 633–657. Allen, F., Qian, J., Qian, M., & Zhao, M. (2009). Review of China’s ﬁnancial system and initiatives for the future. Milken Institute Series on Financial Innovation and Economic Growth, 8,3–72. Bai, J., & Lian, L. S. (2012). Diﬀerence of capital allocation: ownership discrimination or endow- ments deﬁciencies? Management World (Chinese), 6,30–42. Berger, A. N., Klapper, L. F., & Udell, G. F. (2001). The ability of banks to lend to informationally opaque small businesses. Journal of Banking and Finance, 25, 2127–2167. Berger, A., Hasan, I., & Zhou, M. (2009). Institutional development, ownership structure and business strategies: A new era in the Chinese banking industry. Milken Institute Series on Financial Innovation and Economic Growth, 8, 369–402. Bertrand, M., Schoar, A., & Thesmar, D. (2007). Bank deregulation and industry structure: Evidence from the French banking reforms of 1985. Journal of Finance, 62, 597–628. Butler, A., & Cornaggia, J. (2011). Does access to external ﬁnance improve productivity? Evidence from a natural experiment. Journal of Financial Economics, 99, 184–203. Chava, S., Oettl, A., Subramanian, A., & Subramanian, K. (2013). Banking deregulation and innova- tion. Journal of Financial Economics, 109, 759–774. Chen, G., & Weng, W. G. (2013). The eﬀect of foreign banks on the ﬁnancing cost of credit in China. Industrial Economics Research (Chinese), 6,66–76. Chen, W. D., Zong, L., & Zhang, Z. J. (2007). A study for basic characteristics and tendency of foreign banks development: A survey report by questionnaire of foreign banks development. Studies of International Finance (Chinese), 3,38–44. Chen, X. D., Liang, T. Y., Lei, P., & Qin, Q. D. (2015). Inﬂuence of executive compensation incentives on corporate R&D eﬃciency. Science Research Management (Chinese), 9,26–35. Clarke, G., Cull, R., Peria, M. S. M., & Sánchez, S. M. (2005). Bank lending to small businesses in Latin America: Does bank origin matter? Journal of Money Credit & Banking, 37,83–118. Cornaggia, J., Mao, Y. F., Tian, X., & Wolfed, B. (2015). Does banking competition aﬀect innovation? Journal of Financial Economics, 115, 189–209. Degryse, H., Havrylchyk, O., & Jurzyk, E. (2009). Foreign bank entry and credit allocation in emerging markets. European Banking Center Discussion Paper. No.23. Fan, G., Wang, X. L., & Zhu, H. P. (2011). NERI index of marketization for China’s provinces: 2011 Report. Beijing: Economic Science Press. Grosse, R., & Goldberg, L. G. (1991). Foreign bank activity in the United States: An analysis by country of origin. Journal of Banking and Finance, 15, 1092–1112. Hsu, P., Tian, X., & Xu, Y. (2014). Financial development and innovation: Cross country evidence. Journal of Financial Economics, 112, 116–135. 82 BAI ET AL. Jian, Z., Gan, C. H., & Yu, D. F. (2013). Marketization of banking system, allocation of credit and industry restructuring. Economic Research Journal (Chinese), 5, 112–127. Levine, R. (1996). Foreign bank, ﬁnancial development and economic growth. Washington, DC: AEI Press. Levine, R. (2004). The corporate governance of banks: A concise discussion of concept and evidence. World Bank Policy Research Working Paper, No. 3404. Levy, Y. E., & Micco, A. (2007). Concentration and foreign penetration in Latin American banking sectors: Impact on competition and risk. Journal of Banking and Finance, 31, 1633–1647. Li, L., Gao, H. L., & Chen, J. H. (2015). Signaling games for high-tech enterprises bank lending. Economic Research Journal (Chinese), 6, 162–174. Li, W., & Han, L. Y. (2008). A study for foreign bank entry’sinﬂuence on the degree of market competition in banking industry: Empirical study from Panzar-Rosse model. Journal of Financial Research (Chinese), 5,87–98. Li, W. J., & Zheng, M. N. (2016). Is it substantive innovation or strategic innovation? Impact of macroeconomic policies on micro-enterprises’ innovation. Economic Research Journal (Chinese), 4,60–73. Lin, H. D. (2011). Foreign bank entry and ﬁrms’ access to bank credit: Evidence from China. Journal of Banking and Finance, 35, 1000–1009. Lin, Y. F., Sun, X. F., & Jiang, Y. (2009). Toward a theory of optimal ﬁnancial structure in economic development. Economic Research Journal (Chinese), 8,4–17. Liu, P. (2008). Chinese state-owned banking reforms’ system choice. Doctoral Dissertation of Renmin University of China. Mao, Z. S., Wu, J., & Liu, M. L. (2010). An empirical study of foreign banks’ eﬀect on Chinese credit supply. Journal of Financial Research (Chinese), 1, 106–116. Qian, J., Strahan, P. E., & Yang, Z. (2015). The impact of incentives and communication costs on information production: Evidence from bank lending. Journal of Finance, 70, 1457–1495. Stiglitz, J. E., & Andrew, W. (1981). Credit rationing in markets with imperfect information. The American Economic Review, 71, 393–410. Tan, Y., Tian, X., Zhang, C. X., & Zhao, H. (2014). Privatization and innovation: Evidence from a quasi-experiment in China. Kelley School of Business Research Paper, No. 33. Tian, S. H., & Xu, M. D. (2010). Evidences for multinational banks impacts on region source allocation in China. Shanghai Journal of Economics (Chinese), 8,29–40. Tong, T, W., He, W., He, Z. L., & Lu, J. (2014). Patent regime shift and ﬁrm innovation: Evidence from the amendment to China’s patent law. In Academy of Management Proceedings, 1, 14174. Tsai, Y. J., Chen Y. P., Lin C. L., & Hung, J. H. (2014). The eﬀect of banking system reform on investment– cash ﬂow sensitivity: Evidence from China. Journal of Banking and Finance, 46,166–176. Unite, A., & Sullivan, M. (2003). The eﬀect of foreign entry and ownership structure on the Philippine domestic banking market. Journal of Banking and Finance, 27, 2323. Wang, C., & Song, H. Y. (2012). An empirical research on the relationship among the ownership structure, market structure and cost eﬃciency of securities companies in China. Journal of Financial Research (Chinese), 5,80–92. Wen, J., & Feng, G. F. (2012). Heterogeneous institutional investor, nature of ﬁrm and independent innovation. Economic Research Journal (Chinese), 3,53–64. Wu, Y. B. (2007). Firm size, market concentration and innovation: A survey. Economic Research Journal, 5, 125–138. Wu, Y. B., & Mi, Z. Y. (2011). Innovation, imitation and technical eﬃciency in Chinese non-state- owned manufacturing enterprises. Social Sciences in China, 4,77–94. Xie, W. M., & Fang, H. X. (2011). Financial development, ﬁnancing constrains and ﬁrms’ R&D investment. Journal of Financial Research (Chinese), 5, 171–183. Yamori, N. (1998). A note on the location choice of multinational banks: The case of Japanese ﬁnancial institutions. Journal of Banking and Finance, 1, 109–120. Yao, S. J., Feng, G. F., & Jiang, C. X. (2011). Banking reform and eﬃciency in China: 1995-2008. Economic Research Journal (Chinese), 8,4–14. CHINA JOURNAL OF ACCOUNTING STUDIES 83 Yao, Y. J., Wu, W. Q., & Wang, L. L. (2015). Are foreign-invested banks white knights to ease ﬁnancing constraints of Chinese ﬁrms? Empirical study from the perspective of ﬁrm hetero- geneity. Journal of Finance and Economics (Chinese), 10,58–68. Zhang, H. J., & Yang, C. J. (2007). A study for location select and reason of foreign bank entry into China market. Journal of Financial Research (Chinese), 9, 160–172. Zhang, J., Yang, L. X., & Xin, F. (2016). Did real estate hinder the innovation in China? An analysis on loan term structure of ﬁnancial system. Management World (Chinese), 5,64–80. Zhang, J. Q., & Wu, Y. H. (2010). A study for Threshold Eﬀects of foreign bank entry’sinﬂuence on the eﬃciency of domestic commercial bank: Evidence from commercial banks in China. Journal of Financial Research (Chinese), 6,60–74. Zhou, X. P. (2012). Does lending behavior of banks in China vary by ownership? An empirical study based on Markov Regine Switching VAR. Review of Investment Studies (Chinese), 2, 103–112. Zhou, X., Cheng, L. R., &Wang, H. (2012). Is higher the level of technological innovation, better the ﬁnancial performance? An empirical analysis of patent application of the listed Chinese phar- maceutical company. Journal of Financial Research (Chinese), 8, 166–179.
China Journal of Accounting Studies
– Taylor & Francis
Published: Jan 2, 2018
Keywords: banking competition; difference-in-difference approach; enterprise innovation; foreign bank entry