Government intervention and corporate policies: Evidence from China

Government intervention and corporate policies: Evidence from China

JBR-08211; No of Pages 11 Journal of Business Research xxx (2014) xxx–xxx Contents lists available at ScienceDirect Journal of Business Research Go...

502KB Sizes 0 Downloads 92 Views

JBR-08211; No of Pages 11 Journal of Business Research xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

Journal of Business Research

Government intervention and corporate policies: Evidence from China☆,☆☆ Yingying Shao a, Rodrigo Hernández b,1, Pu Liu c,⁎ a b c

Department of Finance, College of Business and Economics, Towson University, USA Department of Accounting, Finance and Business Law, College of Business and Economics, Radford University, USA Department of Finance, Sam M. Walton College of Business, University of Arkansas, USA

a r t i c l e

i n f o

Available online xxxx Keywords: Government intervention Ownership structure Financial structure Investments

a b s t r a c t This study examines two channels through which Chinese government intervenes in business activities: direct intervention via government ownership and indirect intervention via strategic development plans in selected areas. The findings show that these interventions affect corporate policies differently and have opposite effects on financing policies: while firms with higher level of government ownership tend to use higher leverage, more long-term debt and hold less cash, and such effects are more pronounced with central government ownership, reverse effect is related with strategic development plans. In addition, the study shows that indirect intervention alleviates the impact of direct intervention on firms' financing policy. In terms of investment policies, both forms of intervention are related to higher investment expenditures and poorer performance. The effect of government ownership on firms' leverage has become less significant after the establishment of corporate bond market in China. © 2014 Elsevier Inc. All rights reserved.

Introduction Government intervention can have a significant impact on an economy and the effect is particularly profound in emerging countries where the financial markets are more opaque, credit and financing are more difficult to receive, bureaucracies are more severe, and government intervention is more prevalent than in developed countries. In China, after three decades of economic reforms, the government still maintains substantial influences in business activities through different forms of intervention. This paper focuses on two forms of government intervention and examines how they affect corporate financing decisions including financial leverage, the use of long-term debt, and cash holdings. In

☆ We are thankful for the suggestions from the participants in the 2013 Research Conference “Competing in China: Local Firms, Multinationals, and Alliances” sponsored by the George Mason University, School of Management. We are particularly grateful to two anonymous referees, four Guest Editors Robert Grosse, Ning Li, Yan Lin, and Ling Lisic, and the Editor-in-Chief of Journal of Business Research, Arch Woodside for their constructive comments. In addition, we are indebted to Yaru Grace Liu who provided outstanding editing and proofreading of the paper. We thank Jun Duanmu, Lifa Huang, Yongjia Li, and Weineng Xu for their excellent assistance in data collection. The remaining errors, if any, are solely ours. ☆☆ Funding for the research was provided by Harold A. Dulan Professorship in Capital Formation and Robert E. Kennedy Professorship in Investment at Sam M. Walton College of Business, University of Arkansas. ⁎ Corresponding author. Tel.: +1 479 575 6095. E-mail addresses: [email protected] (Y. Shao), [email protected] (R. Hernández), [email protected] (P. Liu). 1 Tel.: +1 540 831 6454.

addition, this study investigates the effect of government intervention on firms' investment decisions and performance, respectively. The first form of intervention is Chinese government's ownership of corporate firms. Studies examining the effect of government ownership have reported that high level of government ownership is often associated with pursuit of political and social objectives (Shleifer & Vishny, 1994), and that state-owned banks often make lending decisions based on social and political goals including providing jobs to the society and/or bailing out financially distressed firms (Cull & Xu, 2003; La Porta, Lopez de Silanes, & Shleifer, 2002). In China, a distinctive feature about government ownership is that the government is often the majority owner of both corporate firms and large banks, also known as dual ownership. The Chinese government owns many large banks including the largest four banks: Industrial and Commercial Bank of China, Agricultural Bank of China, China Construction Bank, and Bank of China, which in total provide more than 80% of commercial and industrial loans to corporations. This provides a unique institutional setting to examine how government ownership affects firms' corporate decisions because dual ownership structure allows the government to instruct the banks it owns to make preferential loans to firms it owns. Even though many of the state-owned enterprises (SOEs) are now partially privatized, they are still under the strong control of the government and still carry the goals of providing social and economic stability in addition to generating profits. The study posits that Chinese firms with higher government ownership may take the advantage of government dual ownership to gain easier access to bank loans. Consequently, these firms tend to have higher leverage, use more long-term debt, and hold less cash. Because these firms can obtain bank loans more easily, they tend to make

http://dx.doi.org/10.1016/j.jbusres.2014.11.015 0148-2963/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

2

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

more capital investment or even undertake value-reducing projects mandated by the government and thus experience relatively poor performance. The results in the study confirm the hypotheses and further reveal that the effect is more pronounced when firms are owned by central government compared to being owned by local government. The second form of government intervention in this study is Chinese government's decision to strategically select some special “economic development areas” (to be referred to as EDAs henceforth). During the economic reforms, China has strategically established four EDAs: Yangtze River Delta Economic Area, Pearl River Delta Economic Area, Beijing–Tianjin–Hebei Economic Area, and Chengdu–Chongqing Economic Area. The government has implemented a number of preferential policies in these areas to create more favorable economic, technical and industrial opportunities so that the developments in these areas can serve as an economic engine to drive the economy in the rest of the nation. While the establishment of these EDAs was successful in boosting the national economy, it also produced unintended consequences such as creating severe economic disparities across regions. According to the China Statistics Yearbook, the average GDP growth rate is about 18% in EDAs compared to 5% in other areas, and the GDP in EDAs represents more than 40% of the total GDP in China. Based on the research suggesting that firms' locations play an important role in corporate decisions (Almazan, Motta, Titman, & Uysal, 2010) and that financial slack is needed to fund potential growth opportunities (Myers and Majluf, 1984), this study conjectures that in China, the economic disparities in different areas caused by the government's development strategies and the resultant variation in growth opportunities would lead firms to make different corporate decisions based on their geographical locations. Such differences in corporate decisions between firms in EDAs and firms in other areas are referred to as the effect of indirect government intervention because the effect is transferred from the government's influence in the entire economy to individual firms' decisions. In other words, the strategic policies implemented by government in EDAs first cause changes in the macroeconomic environment, which consequently lead to changes in firms' behavior. In contrast, government ownership is referred to as direct intervention in the paper. In the literature, few studies have attempted to examine the effect of indirect intervention on corporate decisions, particularly in the form of establishing EDAs. Only recently research began to take a step toward this direction. Chen, Khan, Yu, and Zhang (2013) recognize the difference in regional government interventions and they study the relationship between regional government interventions and firms' co-investments as well as the consequent performance due to firms' investment co-movement. However, it is still unclear how EDA, a different form of government intervention, affects firms' financing decisions, investment decisions, and their performance. The empirical results in this study show that while the direct and indirect interventions have similar effects on firms' investment decisions and performance, their effects on financing decisions are opposite. Firms located in EDAs tend to choose an efficient financial structure as reflected in lower leverage, less use of long-term debt, and more cash holdings than firms in other areas. Also, the investigation on the joint effect of government ownership and locations (in terms of inside or outside EDAs) reveals that the effect of government ownership on financing decisions and performance is less pronounced for firms inside EDAs than that for firms outside EDAs. In sum, this study suggests that while direct government intervention via government ownership may have led to inefficient financing policies and poor performance, indirect government intervention through establishing EDAs plays a positive role in mitigating the adverse effect of government ownership on firm's financing decisions. Lastly the paper examines whether the inception of a new corporate bond market in China in 2007 affects firms' choice of leverage. The participants in bond market, which is free from government intervention,

may behave differently from the state-owned banks. For instance, the lenders in the bond market are more likely to make lending decisions based on borrowers' creditworthiness hence are less willing to lend to poorly performing firms that carry social and political goals. Therefore, we would expect that the inception of the corporate bond market would constrain the ability of government-controlled firms in the increase of financial leverage in order to fulfill the social and political mandates of the government. This paper, to the best of our knowledge, is the first study using geographical EDA locations to examine how government policies affect corporate decisions and performance. The findings on the relationship between government's selection of EDA and firms' corporate decisions suggest a positive role played by the government in inducing efficient financing decisions. Moreover, the finding that indirect government intervention can help mitigate the adverse effect of direct intervention is a major contribution to the literature. The paper also contributes to the literature in the area of government ownership, particularly in the study of China. It provides a comprehensive study on the effect of Chinese government ownership on multiple decisions in financing, investment, and performance, while previous studies tend to focus on the effect of government ownership on one single decision. For instance, Pessarossi and Weill (2013) study how government ownership affects firms' decision on borrowing from bond market vs. syndicated loans. Gul (1999) examines the relationship between government ownership, debt financing, and dividend policies which is only tangentially related to our study in terms of debt financing. In addition, this study complements the existing studies as the findings on the relationship between government ownership and firm performance are largely mixed. Some studies suggest that government ownership may undermine the performance because, in order to fulfill social and political goals, state-owned enterprises may undertake value-reducing investments (Chen, Sun, Tang, & Wu, 2011; Kang & Kim, 2012; Shleifer & Vishny, 1994), while others find that the relationship between government ownership and performance follows an inverted U-shape implying that too little government ownership may not provide SOEs with enough government support during financial distresses, while too much government ownership may lead to too much government interference in operation and management (Sun, Tong, & Tong, 2002). Our study contributes to the literature by suggesting a geographical variation in the effect of government ownership: while government ownership tends to have a negative effect on firm's performance, the negative effect is less pronounced for firms located in EDAs than for firms located in other areas. Also relatively little is known regarding whether central government affects a firm's decisions differently from local (i.e. provincial, city, and county) government. This study therefore contributes to the literature by documenting a more profound effect of ownership by central government than by local government. In addition, it makes a unique contribution by documenting that the establishment of a new corporate bond market in China weakens the effect of government ownership on firms' leverage as the lenders in bond market are prone to lending based on economic as opposed to social and political considerations. The remainder of this study is organized as follows. Section 2 discusses previous studies and develops the hypotheses. Section 3 describes sample selection and data. Section 4 presents the empirical results and Section 5 concludes the paper. Hypothesis development Government ownership, financing and investment decisions Agency theory suggests that firms' financial policies are affected by ownership structure due to potential interest conflicts among various stakeholders. In the context of state-owned firms, studies suggest that, in addition to profitability goals, government-owned enterprises often

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

pursue social and political objectives and the pursuit may reduce the value of the firm (Shleifer & Vishny, 1994). The dual ownership of firms and banks by government provides a unique opportunity to study how Chinese firms' financial decisions are affected. The government, for instance, has an incentive to influence state-owned banks to make commercial loans to financially constrained firms owned by government in order to meet social and political goals. As such, firms with higher level of government ownership tend to have easier access to commercial loans, and therefore tend to increase their financial leverage. With regard to the choice of debt maturity, liquidity risk theory argues that shorter term debt issuers tend to be subject to more monitoring and higher liquidation risk from lenders upon debt renewals (Diamond, 1991). Therefore firms with higher level of government ownership are more likely to borrow longer-term debt from stateowned banks to reduce monitoring and liquidation risk. Based on these arguments, we hypothesize that: H1a. Firms with higher level of government ownership tend to have higher level of leverage and longer term of debt. Another important aspect of financial policy is cash holdings which provide financial flexibility and discretions to firms and a large body of research has documented various firm-specific characteristics in determining an optimal level of cash holdings. These factors include growth opportunities, uncertainty in cash flows, cost of raising funds, dividend payout policy, and access to capital markets (Bates, Kahle, & Stulz, 2009). Drawing from the transaction and precautionary motives for cash holdings, we posit that firms with higher level of government ownership have less incentive to hold cash because they have easier access to bank credit through government's support, leading to a negative relationship between cash holding and firm's government ownership: H1b. Firms with higher level of government ownership tend to hold less cash. With regard to investment decisions, research based on agency theory suggests that investment decisions for firms with large undiversified shareholders may differ from those with dispersed ownership, largely due to the differences in risk preferences, investment horizon, and contracting efficiency with management (Shleifer & Vishny, 1986). Research also shows that government-owned firms often have the social and political mandates (e.g. providing employment opportunities) which tend to induce firms to undertake more investments than based on profitability measures only, sometimes even value-reducing investment projects, leading to poor performance (Chen et al., 2011; Kang & Kim, 2012). We hence hypothesize that: H1c. Firms with higher level of government ownership tend to have higher capital expenditure and poorer performance. Location, financing policy and investment decisions Recent studies suggest that locations have an impact on firms' decisions (Almazan et al., 2010). In addition, research has suggested that firms tend to maintain financial slacks in order to fund potential growth opportunities (Myers and Majluf, 1984). As discussed above, the establishment of EDAs has led to economic disparities across regions in China, and the disparities may lead to different financial decisions made by firms. For instance, firms located in EDAs tend to have more growth opportunities than firms in other areas and therefore firms located in EDAs have a tendency to maintain greater financial slack in order to fund higher growth opportunities. As such, we hypothesize that: H2a. Firms located in EDAs tend to have lower level of debt and shorter debt maturity than firms in other areas.

3

H2b. Firms located in EDAs tend to hold more cash than firms in other areas. With regard to investment decisions, it is likely that the fast growing economy in EDA tends to lead to more growth opportunities. The growth opportunities may lead firms to invest aggressively and even overinvest, which then may lead to firms' underperformance and the following hypothesis: H2c. Firms in EDAs tend to have larger expenditures and underperformance relative to firms in other areas. Data and variable construction The initial sample includes all the Chinese firms listed on the Shanghai Stock Exchange and Shenzhen Stock Exchange between 1991 and 2010. Chinese firms can issue two classes of stocks: Class A shares quoted in Chinese currency and Class B shares quoted in U.S. dollars. This study focuses on firms that issue Class A shares and have data available in Compustat Global. Following the convention in the literature, this study excludes firms in financial and utility industries because firms in these industries are more strictly regulated and tend to have distinctive capital structures. The sample includes firms with financial and accounting data available in Compustat Global. In order to construct the variable of cash flow volatility, firms included in the sample also need to have at least five years of history. The final sample includes 1459 firms. The government ownership data are obtained from Bloomberg and are confirmed through firms' annual filings via Shanghai Stock Exchange's XBRL database and Shenzhen Stock Exchange's CNINF database. We create a dichotomy variable Government ownership to highlight the difference in government ownership across firms. The variable takes the value of one if government ownership is above 30% and zero otherwise (significant government ownership is defined as 30% of ownership or more according to 2006 Guideline of China Securities Regulation Commission). Furthermore, we make a distinction between ownership by local government (at provincial level) and ownership by central government using two distinctive dummy variables: Central government ownership takes the value of one if a firm is controlled by central government and zero otherwise; and Local government ownership takes the value of one if a firm is controlled by local government and zero otherwise. After matching the final sample with a list of entities controlled by central government (released by the central government), there are 179 firms in the sample under the direct control of the central government. The study uses the location of a firm's headquarter in Compustat Global to determine the location of the firm because Chinese firms rarely relocate once incorporated. We use eleven major cities as the representatives of EDAs – Shanghai, Beijing, Chongqing, Shenzhen, Guangzhou, Tianjin, Wuhan, Dongguan, Shenyang, Hangzhou, and Nanjing – the largest cities in the EDAs covering both the east and west regions of China. These cities form the economic engine that drives the entire Chinese economy and have contributed substantially to the regional and national economic growth. We use a dichotomy variable EDA which equals to one if a firm is located in one of the eleven cities, and zero otherwise. Alternatively, we compute a relative location measure using the logarithm of the weighted average of geographic distances (based on longitude and latitude) from the city where a firm is located to major transportation hubs in the EDAs, including Guangzhou, Shanghai, Beijing, Wuhan, Chongqing, Chengdu, and Zhengzhou. By construction, the EDA dichotomy variable and the distance variable highlight the effect of central government's policies on firms' external economic environment. Regional government policies may also have an impact on firms' external economic environment. To ensure that the effect of indirect government intervention from both the central and regional governments is captured, this study includes the National Economic Research Institute Index of Marketization of China

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

4

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

To test firms' choice of financial policies we use the following financial measures constructed from Compustat Global: leverage as measured by the sum of long-term debt and short-term debt divided by the book value of total assets; cash holding as measured by the ratio of cash and short-term investments to the book value of total assets; debt maturity as measured by the ratio of long-term debt to the sum of long-term debt and current liabilities, a measure commonly used in the literature (e.g. Cai, Fairchild, & Guney, 2008; Demirguc-Kunt & Maksimovic, 1999). We recognize the limitation of using long term debt ratio as the sole measure of debt maturity structure. Chinese firms, however, are not required to report debt maturity structure in their filings, we hence are unable to develop alternative measures such as percentage of debts with maturity more than three (five) years, as some studies did using U.S. data. We measure investment decisions using capital expenditure scaled by total assets, and we measure performance using operating income scaled by total assets. Panel A of Table 1 contains the descriptive statistics of government ownership, geographic location, and financial variables used in the

(Fan, Wang, & Zhu, 2011), which reflects the economic reforms at provincial level over time (Wang, Wong, & Xia, 2008). The original index takes a value between 0 and 10 where a higher number indicates a higher degree of marketization for the province. The correlation coefficients of marketization with the EDA dummy and the distance variable are 0.48 and −0.18 respectively and both are statistically significant at 1% level. The results suggest that while the marketization measure and the location variables are correlated to each other, they are not identical. Therefore the marketization index complements the geographical measurements based on EDAs and allows the results of government intervention to be applied in general. The use of the distance measure, given its low correlation with the marketization measure, further enhances the robustness of the results. In the empirical tests, we create a dummy variable which takes the value of one if the marketization index of the province in which a firm is located is above the median and zero otherwise. Chen et al. (2011) suggest that the governments in regions with lower marketization measures tend to impose greater influence on the local economy, which consequently affects the financial decisions of companies in the region. The results in this study support the argument. Table 1 Summary statistics. Panel A. Government intervention and key financial variables

Government ownership EDA Distance Leverage Cash Total assets (in thousands) Profitability Growth Working capital Capital expenditure Cash flow Dividend payout Debt maturity

Mean

Median

St. dev

10th percentile

90th percentile

0.73 0.35 1285 0.23 0.16 4645 0.08 0.12 −0.03 0.06 0.05 0.36 0.23

0 0 1302 0.23 0.13 1268 0.07 0.13 −0.02 0.04 0.04 0 0.12

0.28 0.48 339 0.15 0.13 32,465 0.21 0.46 0.18 0.06 0.06 0.48 0.27

0 0 922 0.02 0.04 357 0.01 −0.21 −0.27 0.01 −0.02 0 0

1 1 1746 0.44 0.33 6320 0.16 0.48 0.21 0.15 0.12 1 0.65

Panel B. Correlation matrix Leverage (1)

(1)

(2)

(3)

(4)

(5)

Debt maturity (2) Cash (3) Government ownership (4) EDA (5) Marketization (6) Assets (7) Profitability (8) Growth (9) Capital expenditure (10)

0.15⁎⁎⁎ −0.37⁎⁎⁎ 0.15⁎⁎⁎ −0.03⁎⁎⁎ −0.07⁎⁎⁎ 0.12⁎⁎⁎ −0.06⁎⁎⁎

1 −0.08⁎⁎⁎ 0.06⁎⁎⁎ −0.07⁎⁎⁎ −0.07⁎⁎⁎ 0.24⁎⁎⁎ 0.05⁎⁎⁎ 0.05⁎⁎⁎ 0.19⁎⁎⁎

1 −0.03⁎⁎⁎ 0.10⁎⁎⁎ 0.17⁎ 0.00⁎⁎⁎ 0.02⁎⁎ 0.05⁎⁎⁎ 0.02⁎⁎

1 0.08⁎⁎⁎ 0.00⁎⁎⁎ 0.14⁎⁎⁎ −0.03⁎⁎⁎ 0.04⁎⁎⁎ 0.07⁎⁎⁎

1 0.48⁎⁎⁎ 0.10⁎⁎⁎ −0.02⁎⁎ −0.02⁎⁎ 0.09⁎⁎⁎

−0.01 0.05⁎⁎⁎

(6)

1 0.31⁎⁎⁎ −0.02⁎ 0.01 0.12⁎⁎⁎

(7)

(8)

(9)

1 0.01⁎⁎⁎ 0.10⁎⁎⁎ 0.22⁎⁎⁎

1 0.28⁎⁎⁎ 0.07⁎⁎⁎

1 0.10⁎⁎⁎

Panel C. Firms with government ownership vs. diffused ownership structure Majority government ownership Mean Leverage Debt maturity Cash

Median

0.24 0.25 0.14

Diffused ownership N

0.23 0.23 0.12

Mean

3461 3461 3461

Gov. vs. diffused

Median

0.23 0.21 0.16

0.22 0.11 0.13

N

t-Value

Z-value

14,066 14,066 14,066

3.40⁎⁎⁎ 9.37⁎⁎⁎ −13.26⁎⁎⁎

1.98⁎⁎ 2.17⁎⁎⁎ −3.58⁎⁎⁎

Panel D. Firms located in EDAs vs. non-EDAs Within EDAs

Leverage Debt maturity Cash Growth Marketization index (level) Government ownership

Non-EDA locations

EDA vs. Non-EDA

Mean

Median

N

Mean

Median

N

t-Value

Z-value

0.23 0.20 0.18 0.14 7.89 0.83

0.22 0.07 0.14 0.14 8.20 1.00

6153 6153 6153 6153 6153 6153

0.24 0.24 0.15 0.12 6.35 0.76

0.23 0.15 0.12 0.12 6.08 1.00

11,374 11,374 11,374 11,374 11,374 11,374

−3.41⁎⁎⁎ −9.37⁎⁎⁎ 13.26⁎⁎⁎ 2.30⁎⁎⁎ 74.35⁎⁎⁎ 10.53⁎⁎⁎

−2.35⁎⁎⁎ −12.62⁎⁎⁎ 9.61⁎⁎⁎ 2.97⁎⁎⁎ 63.40⁎⁎⁎ 10.12⁎⁎⁎

Notes: Descriptive statistics of key variables. See Appendix A for detailed definitions for these variables. ⁎ Represents statisticallysignificant at 1% level. ⁎⁎ Represents statistically significant at 5% level. ⁎⁎⁎ Represents statistically significant at 1% level.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

Fig. 1. Marketization index for economic development areas and non-economic development areas over time.

study. In the sample, about 73% of firms have significant government ownership. This is in line with the observation in Wang et al. (2008) which report that 72% of Chinese firms have government ownership of 20% or higher. About 35% of the firms are located in EDAs and the mean (median) distance between a firm's headquarter and the center of transportation hubs is about 1285 (1302) miles. The average leverage is 23%, similar to the leverage of U.S. firms documented in Harford, Li, and Zhao (2008). On average Chinese firms maintain a cash balance of 16% in the total assets, similar to U.S. firms

5

(Bates et al., 2009). In terms of debt maturity, the average long-term debt to total debt ratio is 0.23, consistent with the study of Cai et al. (2008). Panel B of Table 1 provides the correlation matrix of the key variables. Comparisons of the key financial variables between firms with high level of government ownership and diffused ownership are presented in Panel C of Table 1. Firms with higher level of government ownership tend to have higher leverage, debt with longer-term to maturity, and lower level of cash holdings. In Panel D, the comparisons of key financial variables between firms in EDAs and firms in other areas indicate that firms located in EDAs tend to have lower leverage, debt with shorterterm to maturity, and higher level of cash holding compared to firms located in other areas. It also shows that firms in EDAs (non-EDAs) tend to have a higher (lower) level of sales growth and government ownership. In addition, EDAs are characterized with a higher mean value (7.89) of marketization index relative to other areas (6.35), indicating these two measures are consistent in measuring geographical variation due to regulatory changes. Fig. 1 depicts the marketization index over time for EDAs and non-EDAs. Empirical results Regression results In examining corporate financial policies, the literature suggested a two-stage least square regression to control for the endogenous relationship between leverage and debt maturity (e.g. Barclay, Marx, &

Table 2 Government intervention on firms' leverage. Notes: First stage of simultaneous regression analysis on leverage and debt maturity with dependent variable as the leverage ratio. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. (1)

(2) 0.014⁎⁎⁎ (6.657)

Government ownership Central gov. ownership Local gov. ownership

(3)

0.023⁎⁎⁎ (10.575) 0.009⁎⁎⁎ (4.170)

EDA

(4)

−0.009⁎⁎ (−2.790)

Log of distance

(5)

0.023⁎⁎⁎ (3.532)

(6)

0.009⁎⁎⁎ (4.158) 0.074⁎⁎⁎

0.009⁎⁎⁎ (4.284) 0.073⁎⁎⁎

0.010⁎⁎⁎ (4.511) 0.074⁎⁎⁎

0.009⁎⁎⁎ (4.100) 0.075⁎⁎⁎

0.009⁎⁎⁎ (4.095) 0.074⁎⁎⁎

−0.015⁎⁎⁎ (−5.518) 0.010⁎⁎⁎ (4.702) 0.068⁎⁎

Tax

(2.947) −0.042⁎ (−1.793) −0.316⁎⁎⁎ (−12.773) 0.001⁎⁎⁎

(2.911) −0.041⁎ (−1.791) −0.316⁎⁎⁎ (−12.790) 0.001⁎⁎⁎

(2.955) −0.041⁎ (−1.787) −0.315⁎⁎⁎ (−12.892) 0.001⁎⁎⁎

(2.993) −0.042⁎ (−1.793) −0.316⁎⁎⁎ (−12.713) 0.001⁎⁎⁎

(2.953) −0.041⁎ (−1.807) −0.317⁎⁎⁎ (−12.814) 0.001⁎⁎⁎

(2.662) −0.042⁎ (−1.817) −0.319⁎⁎⁎ (−12.715) 0.001⁎⁎⁎

Financial distress

(3.253) 0.034⁎⁎⁎

Constant

(3.042) −0.020⁎⁎⁎ (−8.523) −0.040⁎⁎⁎ (−4.505) 0.248⁎⁎⁎

(3.283) 0.034⁎⁎⁎ (3.074) −0.018⁎⁎⁎ (−7.415) −0.041⁎⁎⁎ (−4.633) 0.253⁎⁎⁎

(3.286) 0.035⁎⁎⁎ (3.154) −0.018⁎⁎⁎ (−7.427) −0.043⁎⁎⁎ (−4.872) 0.251⁎⁎⁎

(3.231) 0.034⁎⁎⁎ (3.088) −0.021⁎⁎⁎ (−9.544) −0.040⁎⁎⁎ (−4.439) 0.248⁎⁎⁎

(3.331) 0.034⁎⁎⁎ (3.103) −0.021⁎⁎⁎ (−8.940) −0.039⁎⁎⁎ (−4.482) 0.079⁎

(3.409) 0.036⁎⁎⁎ (3.214) −0.020⁎⁎⁎ (−8.438) −0.037⁎⁎⁎ (−4.083) 0.252⁎⁎⁎

Industry fixed effects 2-way cluster analysis Observations R-squared

(8.537) Yes Yes 16,818 0.227

(8.701) Yes Yes 16,818 0.228

(8.437) Yes Yes 16,804 0.229

(8.608) Yes Yes 16,818 0.227

(1.915) Yes Yes 16,818 0.228

(8.663) Yes Yes 16,818 0.229

Marketization Size Asset tangibility Profitability Working capital

Institutional ownership Insider ownership

⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

6

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

Table 3 Government intervention on firms' debt maturity choice. (1)

(2)

(3)

0.023⁎⁎⁎ (4.827)

Government ownership

0.043⁎⁎⁎ (6.490) 0.023⁎⁎⁎ (4.659)

Central gov. ownership Local gov. ownership EDA

(4)

−0.048⁎⁎⁎ (−9.219)

Log of distance

(5)

0.026⁎ (1.884)

Marketization 0.075⁎⁎⁎ (25.692) 0.170⁎⁎⁎

0.075⁎⁎⁎ (25.594) 0.171⁎⁎⁎

0.077⁎⁎⁎ (25.938) 0.174⁎⁎⁎

0.077⁎⁎⁎ (24.882) 0.171⁎⁎⁎

0.074⁎⁎⁎ (25.592) 0.170⁎⁎⁎

Constant

(4.772) −0.094⁎⁎⁎ (−5.028) −0.471⁎⁎⁎ (−6.313) 0.298⁎⁎

(4.811) −0.097⁎⁎⁎ (−5.156) −0.479⁎⁎⁎ (−6.409) 0.306⁎⁎

(4.898) −0.101⁎⁎⁎ (−5.317) −0.483⁎⁎⁎ (−6.466) 0.307⁎⁎

(4.835) −0.095⁎⁎⁎ (−4.930) −0.477⁎⁎⁎ (−6.273) 0.290⁎⁎

Industry fixed effects 2-way cluster analysis Observations R-squared

(2.733) Yes Yes 15,773 0.149

(2.822) Yes Yes 15,773 0.149

(2.828) Yes Yes 15,759 0.150

(2.662) Yes Yes 15,773 0.155

(4.788) −0.093⁎⁎⁎ (−5.059) −0.468⁎⁎⁎ (−6.418) 0.109 (0.556) Yes Yes 15,773 0.149

Size Financial distress Insider ownership Fitted leverage

(6)

−0.069⁎⁎⁎ (−14.008) 0.078⁎⁎⁎ (27.005) 0.174⁎⁎⁎ (5.013) −0.079⁎⁎⁎ (−4.371) −0.501⁎⁎⁎ (−6.392) 0.296⁎⁎ (2.782) Yes Yes 15,773 0.161

Notes: Second stage of simultaneous regression analysis with debt maturity as the dependent variable. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

Smith, 2003; Datta, Iskandar-Datta, & Raman, 2005). The two-stage least square regression framework models leverage and debt maturity simultaneously, with leverage estimated in the first stage and debt maturity structure estimated in the second stage, which also includes fitted leverage estimate as an explanatory variable. The first-stage regression model is: Leveragei;t ¼ α þ β1 Govi þ β2 X i;t þ β3 Industryi þ εi;t ;

ð1Þ

where Govi measures government interventions in two forms as described before. Xi,t is a vector of control variables including: firm size, asset tangibility, profitability, growth opportunities, working capital, capital investment, effective tax rate, financial distress, insider ownership, institutional ownership, and stock market volatility. Industryi is a vector of industry dummy variables defined by the Fama and French (1997) classification. Given that the data is an unbalanced panel, we employ the panel data estimator with a two-way cluster control at the time and firm levels (Petersen, 2009). The definition and measurement of each of these variables are described in the Appendix of the paper. In order to mitigate the impact of outliers, we winsorize all variables at the 1% level at both tails. The second-stage regression for debt maturity is specified as follows: DebtMaturityi;t ¼ α þ β1 Govi þ β2 FittedLeveragei;t þ β3 Y i;t þ β4 Industryi þ εi;t ;

ð2Þ

where Fitted leverage is the predicted leverage from the first-stage regression. To properly identify the system of equations, the second stage model excludes asset tangibility and profitability. The results of the first stage regression on leverage are in Table 2. Column (1) shows that the effects of control variables are broadly consistent with earlier studies on determinants of financial leverage. For example, larger firms, higher tangible assets, lower profitability,

lower net working capital, and higher effective tax rate tend to be related to higher level of leverage. The negative relationships between leverage and institutional ownership as well as insider ownership are consistent with the agency theory that suggests firms with higher insider ownership tend to choose lower level of leverage to minimize monitoring by lenders, and institutional owners tend to be informed investors thus they are often substitutes for lenders in monitoring management (Shleifer & Vishny, 1986). Consistent with the hypothesis on firms' leverage and the study of Gul (1999), the results in Column (2) show a significant and positive relationship between government ownership and leverage. The results in Column (3) indicate that central government ownership has a stronger impact on leverage than local government ownership. The results in Columns (4) and (5) suggest that firms located in EDAs tend to use lower leverage compared to firms in other areas, consistent with the hypothesis that firms located in EDAs have more financial slack for potential investment opportunities available in these areas. Comparing Column (2) with Column (4), the results suggest that direct and indirect government interventions have opposite effects on firms' leverage. In Column (6), the significantly positive coefficient for the marketization index is consistent with the hypothesis that firms located in areas with better institutional quality have less need for debt financing. Table 3 reports the results for the second stage regressions on firms' debt maturity structure. The results in Column (1) are in line with the findings in the literature that larger firms, firms with lower leverage, and firms in financial distress tend to use more long-term debt. The significant and positive relationship between government ownership and debt maturity structure, as shown in Column (2), is consistent with the hypothesis that firms with higher government ownership tend to use more long-term debt. The results in Column (3) suggest that central government ownership has a more pronounced effect on the use of long-term debt than local government ownership. The results in

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

Columns (4) and (5) show that firms located in EDAs tend to have shorter debt maturity, consistent with the hypothesis that in order to finance growth opportunities, firms tend to use short-term debt to avoid underinvestment problem associated with long-term borrowing. Comparing Column (2) with Column (4), the results suggest that direct and indirect government interventions have opposite effects on the use of long-term debt. Lastly, Column (6) shows that firms located in areas with higher quality of institutions tend to use less long-term debt. This finding is similar to the findings on non-publicly traded firms in China by Li, Yue, and Zhao (2009), who argue that the less use of long-term debt is due to the availability of alternative financing sources resulting from China's economic reforms. The examination on firms' cash holding uses the following specification: Cashi;t ¼ α þ β1 Govi þ β2 Cashi;t−1 þ β3 Y i;t þ β4 Industryi þ εi;t :

ð3Þ

Since cash balance can be viewed as negative debt, the effect of government intervention on cash balance is expected to be opposite to the effect on leverage. Following the literature, Eq. (3) includes a one-period lagged cash ratio to control for the persistence in cash holdings. Other control variables included in vector Yi,t are described in the Appendix. The results in Column (1) of Table 4 are similar to early studies on cash balance held by U.S. firms: larger firms, firms with lower capital

7

expenditure and lower working capital tend to hold more cash. The results in Column (2) show a significant and negative relationship between firms' cash holding and government ownership. State-owned firms have less incentives to hold extra cash because they have easier access to credit from government-owned banks in case they need cash. The results in Column (3) show that government ownership on firms' cash holdings is more pronounced for firms with central government ownership than local government ownership, because in China central government has a greater influence on state-owned banks than local government. The results in Columns (4) and (5) indicate that firms located in or close to EDAs tend to hold more cash as they tend to have more growth opportunities than firms in other areas. Overall, the results once again indicate that direct and indirect government interventions have opposite effects on firm's cash holdings. The results of examining the effect of government intervention on firms' investment decisions and performance are reported in Tables 5 and 6 respectively. Table 5 provides strong evidence that government intervention, either in direct form (through government ownership) or in indirect form (through establishment of EDAs) has a significantly positive impact on firms' capital investment. The positive relationship between government ownership and capital investment suggests that firms with higher government ownership may take advantage of their easier access to banks' credit to invest aggressively, therefore, they tend to have higher capital investment. The positive coefficient of EDA suggests that firms in EDAs, driven by growth opportunities, tend

Table 4 Government intervention on firms' cash holdings. (1)

(2) −0.001⁎ (−1.988)

Government ownership Central gov. ownership Local gov. ownership

(3)

−0.002⁎⁎ (−2.988) 0.002 (0.988)

EDA

(4)

0.003⁎⁎ (2.159)

Log of distance

(5)

(6)

−0.010⁎⁎⁎ (−4.881)

Marketization

0.002 (1.201) 0.641⁎⁎⁎

0.641⁎⁎⁎ (22.243) −0.004⁎⁎ (−2.183) 0.111⁎⁎⁎ (4.353) 0.002⁎

0.641⁎⁎⁎ (22.237) −0.004⁎⁎ (−2.189) 0.111⁎⁎⁎ (4.346) 0.002⁎

0.641⁎⁎⁎ (22.373) −0.004⁎⁎ (−2.209) 0.110⁎⁎⁎ (4.389) 0.002⁎

0.640⁎⁎⁎ (22.198) −0.004⁎⁎ (−2.201) 0.112⁎⁎⁎ (4.349) 0.002⁎

0.640⁎⁎⁎ (22.390) −0.004⁎⁎ (−2.257) 0.109⁎⁎⁎ (4.277) 0.002⁎

(22.430) −0.004⁎⁎ (−2.278) 0.111⁎⁎⁎ (4.358) 0.002⁎

Leverage

(1.920) −0.078⁎⁎⁎ (−7.871) −0.220⁎⁎⁎ (−7.196) −0.156⁎⁎⁎

(1.895) −0.078⁎⁎⁎ (−7.860) −0.220⁎⁎⁎ (−7.183) −0.156⁎⁎⁎

(2.073) −0.078⁎⁎⁎ (−7.885) −0.220⁎⁎⁎ (−7.197) −0.156⁎⁎⁎

(1.761) −0.078⁎⁎⁎ (−7.856) −0.218⁎⁎⁎ (−7.125) −0.156⁎⁎⁎

(1.970) −0.077⁎⁎⁎ (−7.809) −0.221⁎⁎⁎ (−7.254) −0.156⁎⁎⁎

(1.791) −0.078⁎⁎⁎ (−7.865) −0.219⁎⁎⁎ (−7.165) −0.156⁎⁎⁎

Dividend payout

(−10.477) 0.009⁎

(−10.456) 0.009 (1.720) 0.020⁎ (1.917) 0.141⁎⁎⁎ (3.705) Yes Yes 15,478 0.536

(−10.593) 0.009 (1.727) 0.020⁎ (1.972) 0.141⁎⁎⁎ (3.706) Yes Yes 15,465 0.536

(−10.408) 0.009 (1.711) 0.020⁎ (1.907) 0.142⁎⁎⁎ (3.725) Yes Yes 15,478 0.536

(−10.624) 0.009 (1.718) 0.020⁎ (1.925) 0.213⁎⁎⁎ (5.261) Yes Yes 15,478 0.536

(−10.394) 0.009 (1.701) 0.020⁎ (1.903) 0.141⁎⁎⁎ (3.721) Yes Yes 15,478 0.536

Lagged cash Cash flow risk Cash flow Size Working capital Capital expenditure

Financial distress Constant Industry fixed effects 2-way cluster analysis Observations R-squared

(1.735) 0.020⁎ (1.920) 0.141⁎⁎⁎ (3.723) Yes Yes 15,478 0.536

Notes: Regression analysis on firm's cash holding with cash ratio as the dependent variable. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

8

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

Table 5 Government intervention on firms' investments. (1)

(2) 0.002⁎⁎ (2.194)

Government ownership Central gov. ownership Local gov. ownership

(3)

0.007⁎⁎⁎ (5.446) 0.003⁎⁎⁎ (3.291)

EDA

(4)

0.013⁎⁎⁎ (6.817)

Log of distance

(5)

−0.012⁎⁎⁎ (−4.657)

(6)

0.010⁎⁎⁎ (8.685) 0.009⁎⁎⁎

0.010⁎⁎⁎ (8.736) 0.009⁎⁎⁎

0.010⁎⁎⁎ (9.016) 0.009⁎⁎⁎

0.011⁎⁎⁎ (9.047) 0.009⁎⁎⁎

0.010⁎⁎⁎ (8.657) 0.009⁎⁎⁎

0.005⁎⁎ (2.603) 0.010⁎⁎⁎ (8.766) 0.009⁎⁎⁎

Financial distress

(7.090) −0.036⁎⁎⁎ (−5.276) 0.014⁎⁎⁎ (2.897) 0.020⁎

(7.099) −0.036⁎⁎⁎ (−5.272) 0.014⁎⁎⁎ (2.905) 0.020⁎

(7.064) −0.035⁎⁎⁎ (−5.202) 0.014⁎⁎⁎ (2.940) 0.020⁎

(6.793) −0.037⁎⁎⁎ (−5.415) 0.014⁎⁎⁎ (3.068) 0.020⁎

(6.992) −0.035⁎⁎⁎ (−5.188) 0.013⁎⁎⁎ (2.908) 0.020⁎

(6.904) −0.036⁎⁎⁎ (−5.295) 0.014⁎⁎⁎ (2.927) 0.020⁎

Insider ownership

(1.851) 0.029⁎⁎⁎

(1.857) 0.029⁎⁎⁎ (4.640) 0.043 (1.015) Yes Yes 15,478 0.125

(1.909) 0.028⁎⁎⁎ (4.518) 0.043 (1.021) Yes Yes 15,465 0.126

(1.891) 0.030⁎⁎⁎ (4.793) 0.039 (0.946) Yes Yes 15,478 0.134

(1.857) 0.029⁎⁎⁎ (4.642) 0.127⁎⁎ (2.831) Yes Yes 15,478 0.127

(1.871) 0.030⁎⁎⁎ (4.791) 0.042 (0.990) Yes Yes 15,478 0.126

Marketization Size Growth Working capital Dividend payout

Constant Industry fixed effects 2-way cluster analysis Observations R-squared

(4.657) 0.042 (1.001) Yes Yes 15,478 0.125

Notes: Regression analysis on firm's investment with capital expenditure ratio as the dependent variable. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

to investment more. Overall, the results show that both forms of government interventions promote investment. The results in Table 6 show that government intervention is negatively related to firms' performance, similar to the results reported in the literature (Chen et al., 2011; Li et al., 2009). Relating this finding to the results in Table 5, firms located in EDA are more likely to expand aggressively in those areas, they may undertake risky investments which lead to poorer performance. The results in the study so far have shown that direct and indirect government interventions affect corporate financing policies differently but have same impact on investment decisions and performance. To further explore the joint effect of these two forms of intervention, we separate the sample into two subsamples, one includes firms located inside EDAs only and the other includes firms located outside EDAs only, and then run regressions for each subsample. For the sake of brevity, Table 7 only reports the regression coefficients for government ownership variables. The results for firms located inside EDAS are presented in Panel A and the results for firms outside EDAs are presented in Panel B. The results show that in terms of financing policy, the effects of government ownership for firms located outside EDAs are more pronounced than firms located inside EDAs. This finding, which is new to the literature, suggests that one form of government intervention (establishment of EDAs) could partially reduce the adverse effect of another form of intervention (government ownership) in firms' financing policy. One may also find that government intervention through ownership has a strong positive impact on firms' investment regardless of firms' locations, while the negative effect on performance is only observed in firms located in non-EDAs. This is probably because governments in these areas have a tendency requiring firms under their

jurisdictions to overinvest, even in less profitable or non-profitable projects, in order to meet social and political objectives. In the last test we investigate how the newly created corporate bond market affects firms' capital structure as the participants in the bond market may behave differently from state-owned banks. Bond market participants are more likely to make decisions based on economic factors as opposed to government social and political objectives, and with less influences from the government. We augment our leverage and maturity baseline equations with a differences-in-difference technique to examine whether the public debt market have a different effect on firms' financial policy after 2007. Specifically, in Eqs. (1) and (2), we add two variables: a dummy variable Bond2007 which takes the value of one for years after 2007 to indicate the inception of the corporate bond market, and zero otherwise; and an interaction term between Bond2007 and government ownership. As such, the test is able to examine if the new funding source from the debt market has weakened the effect of government intervention on leverage and debt maturity. The coefficients of the interaction term in Columns (1) and (4) in Table 8 suggest that the availability of funding from the new public debt market reduced the effect of government ownership on firms' leverage but not on debt structure. We also repeat the analyses by separating the entire samples into two subsamples: firms issuing bonds and firms not issuing bonds. We manually collect the information about whether a firm issued bonds or not from Bloomberg and identified 445 (out of the 1459 firms) that issued bonds in the market. The results in Columns (2) and (3) show that for firms issuing bonds, the effect of government ownership on leverage is significantly reduced after 2007, suggesting the new corporate bond market weakened the influence of government in the buildup of firms' leverage.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

9

Table 6 Government intervention on firms' performance. (1)

(2)

(3)

−0.002⁎ (−2.073)

Government ownership Central gov. ownership Local gov. ownership

−0.008⁎⁎ (−2.748) −0.003 (−1.576)

EDA

(4)

−0.011⁎⁎⁎ (−5.120)

Log of distance

(5)

−0.018⁎⁎⁎ (−6.718)

Marketization Size

0.014⁎⁎⁎

0.014⁎⁎⁎

Growth

(9.549) 0.051⁎⁎⁎

Financial distress

(6.005) 0.028⁎⁎ (2.452) −0.087⁎⁎⁎ (−10.828) 0.059⁎⁎⁎

Institutional ownership

(3.964) 0.005⁎⁎⁎

Working capital Leverage

(5.404) 0.017⁎⁎⁎ (3.198) 0.162⁎⁎⁎ (2.974) Yes Yes 15,445 0.185

Insider ownership Constant Industry fixed effects 2-way cluster analysis Observations R-squared

(9.444) 0.051⁎⁎⁎ (6.002) 0.028⁎⁎ (2.445) −0.087⁎⁎⁎ (−10.777) 0.059⁎⁎⁎

0.014⁎⁎⁎ (9.776) 0.051⁎⁎⁎

0.014⁎⁎⁎ (10.030) 0.051⁎⁎⁎

(5.995) 0.028⁎⁎ (2.448) −0.087⁎⁎⁎ (−10.932) 0.060⁎⁎⁎

(3.967) 0.006⁎⁎⁎ (5.564) 0.017⁎⁎⁎ (3.168) 0.163⁎⁎⁎ (2.990) Yes Yes 15,445 0.185

(4.029) 0.006⁎⁎⁎ (5.908) 0.016⁎⁎⁎ (3.073) 0.163⁎⁎⁎ (2.985) Yes Yes 15,432 0.185

(6)

0.014⁎⁎⁎

0.000 (0.045) 0.014⁎⁎⁎ (9.768) 0.051⁎⁎⁎

(5.966) 0.027⁎⁎ (2.358) −0.087⁎⁎⁎ (−11.219) 0.059⁎⁎⁎

(9.499) 0.051⁎⁎⁎ (5.977) 0.030⁎⁎ (2.557) −0.085⁎⁎⁎ (−11.024) 0.059⁎⁎⁎

(6.011) 0.028⁎⁎ (2.456) −0.087⁎⁎⁎ (−11.093) 0.059⁎⁎⁎

(3.957) 0.006⁎⁎⁎ (5.855) 0.018⁎⁎⁎ (3.284) 0.160⁎⁎⁎ (2.910) Yes Yes 15,445 0.188

(3.920) 0.006⁎⁎⁎ (5.877) 0.016⁎⁎⁎ (3.028) 0.289⁎⁎⁎ (4.479) Yes Yes 15,445 0.187

(3.959) 0.005⁎⁎⁎ (5.314) 0.017⁎⁎⁎ (3.182) 0.162⁎⁎⁎ (2.974) Yes Yes 15,445 0.185

Notes: Regression analysis on firm's performance with operating income scaled by the total assets as the dependent variable. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

Predicative validity of the models In a recent study, Woodside (2013) argues for the importance of examining the predictive validity of regression models in empirical data

Table 7 Joint effect of direct and indirect government intervention. Leverage

Cash

Investment

Performance

Panel A: Firms located in EDAs Government 0.004 0.027⁎⁎ ownership (0.152) (2.425) Central gov. 0.031⁎⁎⁎ 0.032⁎⁎

−0.001 (−1.479) −0.002⁎

0.002⁎⁎ (2.311) 0.005⁎⁎

ownership Local gov. ownership

(−1.796) 0.000 (0.023)

(2.132) 0.003⁎ (1.881)

0.002 (0.789) 0.002 (0.452) −0.002 (−1.059)

−0.001⁎⁎ (−2.146) −0.003⁎⁎ (−2.156) 0.002 (0.803)

0.002⁎⁎ (2.628) 0.006⁎⁎⁎ (2.953) 0.003⁎⁎ (2.160)

−0.002⁎ (−1.880) −0.009⁎⁎ (−2.455) −0.002 (−0.776)

(3.950) 0.004 (0.314)

Debt maturity

(2.149) 0.026⁎⁎⁎ (3.032)

Panel B: Firms located outside EDAs Government 0.018⁎⁎⁎ 0.022⁎⁎⁎ ownership (9.850) (3.877) Central gov. 0.024⁎⁎⁎ 0.031⁎⁎⁎ ownership (5.479) (2.756) Local gov. 0.013⁎⁎⁎ 0.021⁎⁎⁎ ownership (7.248) (4.011)

Notes: Subsample regression analyses for the joint effect of direct and indirect forms of government ownership on corporate policies. Only the regression coefficients for government ownership variables are reported for the sake of brevity. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

analysis. Although multiple regression analysis is of great value in revealing significant relationship among economic variables, one needs to be cautious about the validity of predictive model in implementing this technique. Along this line of thought, we examine the predictive validity of multiple regression models used in this study, specifically, the 2-stage least square regression on leverage and debt maturity, the cash regression model, the investment and performance model, using a random sample approach described as follows: First, we split the full sample into two subsamples and run model with one subsample. The regression coefficients based on this subsample are then applied to the other subsample to predict the dependent variable. The predicted values of the dependent variable for the second subsample are then compared to the actual values by examining the correlation coefficients, which are reported in Panel A of Table 9. We repeat the above procedure with a reserve order of subsamples and report the results in Panel B. For the sake of brevity, Panel A and Panel B only list the correlation coefficients between the predicted values and actual values of dependent variables, without reporting the regression coefficients of independent variables in various regression models. Woodside (2013) suggests that under such algorithm, a significant correlation between predicted and actual dependent values reveals acceptable predictive validity of the models. In Table 9, the significant correlation coefficients in both ways of sample splitting suggest that the models adopted in this paper provide sufficient predictive validity. Conclusion This paper provides a comprehensive study on how Chinese firms' financial decisions are affected by two forms of government intervention: direct intervention in the form of government ownership and

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

10

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

Table 8 Effect of newly developed corporate bond markets. Leverage

Government ownership Bond2007 Bond2007 × gov. own

(2)

(3)

(4)

(5)

(6)

Full

Bond

Non-bond

Full

Bond

Non-bond

0.016⁎⁎⁎

0.014⁎⁎

(5.515) −0.050⁎⁎⁎ (−6.719) −0.006⁎⁎

(2.698) −0.013⁎

0.024⁎⁎⁎ (7.755) −0.069⁎⁎⁎

0.024⁎⁎⁎ (3.801) −0.010⁎⁎

0.024⁎⁎⁎ (4.481) −0.020⁎

(−1.985) −0.019⁎⁎⁎ (−3.965) 0.008⁎⁎⁎ (5.156) 0.019⁎ (1.893) −0.009⁎⁎

(−7.388) 0.008⁎⁎ (2.156) 0.010⁎⁎⁎ (3.312) 0.002 (0.151) −0.024⁎⁎⁎

(−1.985) 0.004 (0.495) 0.067⁎⁎⁎ (19.444) 0.154⁎⁎⁎ (3.466) −0.009⁎⁎

(−2.097) 0.002 (0.201)

(−11.910) −0.025⁎⁎⁎ (−4.634)

(−2.625) 0.007 (1.258) 0.077⁎⁎⁎ (29.656) 0.159⁎⁎⁎ (4.185) 0.005 (1.357) −0.091⁎⁎⁎

0.025 (1.667) −0.001 (−0.061) 0.014 (0.785) 0.076⁎⁎⁎ (28.786) 0.122⁎⁎⁎ (3.472) 0.024⁎⁎⁎

0.103⁎⁎⁎ (6.519) −0.371⁎⁎⁎

0.060⁎⁎ (2.712) −0.028⁎

(−4.276) −0.556⁎⁎⁎ (−7.409)

(4.025) −0.102⁎⁎⁎ (−4.911) −0.384⁎⁎⁎ (−5.280)

(−2.223) −0.093⁎⁎⁎ (−3.416) −0.633⁎⁎⁎ (−7.570)

(−1.783) −0.313⁎⁎⁎ (−13.678) −0.169⁎⁎⁎ (−2.973) 0.001⁎⁎⁎ 0.280⁎⁎ (2.427) Yes Yes 15,773 0.154

0.036 (0.327) Yes Yes 5021 0.199

0.381⁎⁎⁎ (3.009) Yes Yes 10,752 0.104

Institutional ownership

(−2.174) 0.013⁎⁎⁎ (4.962) −0.326⁎⁎⁎ (−14.166) −0.113⁎

Insider ownership

(−2.069) 0.001⁎⁎⁎

Size Financial distress

Debt maturity

(1)

(3.363) Fitted leverage Asset tangibility

Tax

0.008 (0.869) −0.019⁎⁎⁎ (−7.128) −0.017⁎⁎⁎ (−3.101) 0.068⁎⁎⁎ (3.477) −0.038⁎

Constant

(−1.754) 0.171⁎⁎⁎

(−5.451) −0.338⁎⁎⁎ (−13.831) −0.017 (−0.350) 0.006 (1.461) 0.214⁎⁎⁎

Industry fixed effects 2-way cluster analysis Observations R-squared

(5.726) Yes Yes 16,818 0.245

(5.294) Yes Yes 5212 0.278

Profitability Working capital Capital expenditure

(4.821) 0.181⁎⁎⁎ (5.506) Yes Yes 11,606 0.258

Notes: Simultaneous regression analyses for the impact of newly developed corporate bond markets on leverage and debt maturity. See Appendix A for detailed definitions of other variables. All regressions include Fama and French (1997) industry fixed effects. Standard errors are clusters at both year and firm levels. The t-statistics are shown in parentheses. ⁎⁎⁎ Denotes significance at the 1% level. ⁎⁎ Denotes significance at the 5% level. ⁎ Denotes significance at the 10% level.

indirect intervention in the form of government's strategic selection of geographical areas for special development. The results in the paper show that these two forms of intervention affect firms' financing decisions differently: direct intervention in the form of government ownership is related to higher level of leverage, more use of long-term debt, and less cash holdings, while indirect intervention has the opposite effect. In addition, both forms of intervention are related to higher level of capital investment and poorer performance. Moreover, ownership by the central government has a more pronounced effect on corporate policies than local government.

This study also reveals a geographical variation in the effect of government ownership on corporate decisions: such effect is less pronounced in government's strategically selected areas than in other areas. The results suggest that establishment of EDAs (indirect government intervention) could have mitigated the adverse effect of government ownership (direct government intervention) particularly on firms' financing decisions. Further evidence indicates that the additional credit source from a newly established corporate bond market reduces the effect of government ownership on firms' leverage.

Table 9 Robustness check: Predictive validity tests of the regression models.

Panel A: using the 1st half of sample to predict the 2nd half of sample Correlation between predicted value and actual value of dependent variable

Panel B: using the 2nd half of sample to predict the 1st half of sample Correlation between predicted value and actual value of dependent variable

Leverage and Debt 2-stage regression (p-value)

Cash Regression (p-value)

Investment Regression (p-value)

Performance Regression (p-value)

0.2917 (b0.0001)

0.749 b0.0003

0.2498 b0.0004

0.3713 b0.0005

0.4319 b0.0001

0.7001 b0.0003

0.2709 b0.0004

0.3841 b0.0005

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015

Y. Shao et al. / Journal of Business Research xxx (2014) xxx–xxx

11

Appendix A. Variable definitions

Variables

Definitions

Government ownership EDA

A dummy variable taking the value of one if the government owns more than 30% of the share of stock, and zero otherwise. A dummy variable taking value of one if a firm's headquarter is located in an economic development area (EDA, or one of the largest eleven cities in China), and zero otherwise. Weighted average of geographic distances from a firm's location to seven transportation hubs in China, based on longitude and latitude. An index value between 0 and 10, with higher value indicating higher degree of marketization in a province in China. A dummy variable equals to one if the marketization index of a province in which a firm is located is above the median and zero otherwise. Long term debt plus debt in current liabilities divided by total assets Cash and short term investments divided by total assets Natural logarithm of book value of total assets Net property, plant and equipment divided by total assets Operating income before depreciations divided by total assets Growth rate in sales Net working capital divided by total assets Capital expenditures to total assets Effective tax rate, measured as tax expenses divided by pretax income Standard deviation of cash flow in past five years Earnings after interest, dividends, and taxes but before depreciation divided by total assets Dummy variable taking on the value 1 if the company pays dividends and zero otherwise Long term debt divided by total debt An index computed as (−0.737 ∗ Size) + (0.043 ∗ Size2) − (0.040 ∗ Age), following Hadlock and Pierce (2010) A dummy variable taking the value of one for years after the launch of corporate bond market and zero otherwise

Distance Marketization (level) Marketization (Dummy) Leverage Cash Size Asset tangibility Profitability Growth Working capital Capital expenditure Tax Cash flow risk Cash flow Dividend payout Debt maturity Financial distress Bond2007

References Almazan, A., Motta, A. D., Titman, S., & Uysal, V. (2010). Financial structure, liquidity and firm locations. Journal of Finance, 65, 529–563. Barclay, M. J., Marx, L. M., & Smith, C. W. (2003). The joint determination of leverage and maturity. Journal of Corporate Finance, 9, 149–167. Bates, T. W., Kahle, K. M., & Stulz, R. M. (2009). Why do U.S. firms hold so much more cash than they used to? Journal of Finance, 64, 1985–2021. Cai, K., Fairchild, R., & Guney, Y. (2008). Debt maturity structure of Chinese companies. Pacific-Basin Finance Journal, 16, 268–297. Chen, D., Khan, S., Yu, X., & Zhang, Z. (2013). Government intervention and investment comovement: Chinese evidence. Journal of Business Finance & Accounting, 40, 564–587. Chen, S., Sun, Z., Tang, S., & Wu, D. (2011). Government intervention and investment efficiency: Evidence from China. Journal of Corporate Finance, 17, 259–271. Cull, R., & Xu, L. C. (2003). Who gets credits? The behavior of bureaucrats and state banks in allocating credit to Chinese state-owned enterprises. Journal of Development Economics, 71, 533–559. Datta, S., Iskandar-Datta, M., & Raman, K. (2005). Managerial stock ownership and the maturity structure of corporate debt. Journal of Finance, 60, 2333–2350. Demirguc-Kunt, A., & Maksimovic, V. (1999). Institutions, financial markets, and firm debt maturity. Journal of Financial Economics, 54, 295–336. Diamond, D. W. (1991). Debt maturity structure and liquidity risk. Quarterly Journal of Economics, 106, 709–737. Fama, E. F., & French, K. R. (1997). Industry costs of equity. Journal of Financial Economics, 43, 153–193. Fan, G., Wang, X., & Zhu, H. (2011). The report on the relative process of marketization of regions in China. Beijing: The Economic Science Press (in Chinese). Gul, F. A. (1999). Government share ownership, investment opportunity set and corporate policy choices in China. Pacific-Basin Finance Journal, 7, 157–172.

Hadlock, C. J., & Pierce, J. R. (2010). New evidence on measuring financial constraints: Moving beyond the KZ index. Review of Financial Studies, 23, 1909–1940. Harford, J., Li, K., & Zhao, X. (2008). Corporate boards and the leverage and debt maturity choices. International Journal of Corporate Governance, 1, 3–27. Kang, Y., & Kim, B. (2012). Ownership structure and firm performance: Evidence from Chinese corporate reform. China Economic Review, 23, 471–481. La Porta, L., Lopez de Silanes, F., & Shleifer, A. (2002). Government ownership of banks. Journal of Finance, 57, 265–301. Li, K., Yue, H., & Zhao, L. (2009). Ownership, institutions, and capital structure: Evidence from China. Journal of Comparative Economics, 37, 471–490. Myers, S., & Majluf, N. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 5, 187–221. Pessarossi, P., & Weill, L. (2013). Choice of corporate debt in China: The role of state ownership. China Economic Review, 26, 1–16. Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies, 22. Shleifer, A., & Vishny, R. W. (1986). Large shareholders and corporate control. Journal of Political Economy, 94, 461–488. Shleifer, A., & Vishny, R. W. (1994). Politicians and firms. Quarterly Journal of Economics, 109, 995–1025. Sun, Q., Tong, W. H. S., & Tong, J. (2002). How does government ownership affect firm performance? Evidence from China's privatization experience. Journal of Business Finance and Accounting, 29, 1–27. Wang, Q., Wong, T. J., & Xia, L. (2008). State ownership, institutional environment, and auditor choice: Evidence from China. Journal of Accounting and Economics, 46, 112–134. Woodside, A. G. (2013). Moving beyond multiple regression analysis to algorithms: Calling for adoption of a paradigm shift from systematic to asymmetric thinking in data analysis and crafting theory. Journal of Business Research, 66, 463–472.

Please cite this article as: Shao, Y., et al., Government intervention and corporate policies: Evidence from China, Journal of Business Research (2014), http://dx.doi.org/10.1016/j.jbusres.2014.11.015