Returnee Directors and Corporate Tax Avoidance

Abstract

This study evaluates the effect of returnee directors on corporate tax avoidance by using data on publicly listed Chinese companies from 2000 to 2012. Returnee directors grow up in China and then study or work abroad before returning home to be listed firms’ board directors. We use the introduction of provincial policies toward attracting skilled individuals with foreign experience as an instrumental variable for Returnee directors, which is the fraction of returnee directors divided by the total number of directors within a firm. Using quantile regression, we find a positive relation between Returnee directors and corporate tax avoidance for low levels of tax avoidance but a negative relation for high levels of tax avoidance. The result is robust to a battery of tests. The relation between returnee directors and tax avoidance is stronger for state-owned enterprises (SOEs) than non-SOEs and stronger for returnees who hold MBA degrees, possess a background in accounting or auditing, or are independent directors than other returnees.

Keywords

corporate tax avoidance returnee directors state-owned enterprises director ability director independence

Introduction

In this study, we evaluate how returnee directors affect corporate tax avoidance by using data on publicly listed Chinese companies from 2000 to 2012. Specifically, we consider Chinese directors who have gained experience overseas and have subsequently returned to China. By examining how firms’ tax avoidance practices are related to the fraction of returnee directors serving on the board, we provide insights into how corporate tax avoidance practices change in an emerging country.

We view tax avoidance as one of the many activities available to a firm for increasing profits. We examine how returnee directors affect tax avoidance for two reasons. First, tax avoidance strategies can be more opaque than other operating activities and hence, may impose unnecessary risk on shareholders. Second, tax avoidance may substitute for other operating activities that may require more costly effort from managers and directors. Examining how returnee directors affect tax avoidance strategies may uncover potentially rich insights on the operating environment of the firm.

Returnee directors, who grow up in China and then study or work abroad before they come back home to be listed firms’ board directors, can help resolve listed firms’ agency problems in emerging markets for two reasons. First, given their expertise that could help monitor corporations and advise firms, these returnee directors can effectively monitor the firms’ management. Monitoring and advising can alleviate agency issues related to tax avoidance and lead to more efficient tax practices. Second, given their relatively greater independence compared with local directors or managers, as well as their weaker attachment to local politicians or other local parties, returnee directors can have stronger incentives as monitors and advisors.

China offers a unique setting for studying how directors with foreign experience can change tax avoidance. First, Chinese listed firms demand managers and directors who can conduct business in an international environment. As there is a general shortage of directors with foreign experience, there is a wide cross-firm variation in their representation on boards. Second, returnee directors have work experience in different countries and diverse personal characteristics, which allow us to examine the features that relate most strongly to tax avoidance. Third, most Chinese provinces introduced policies that attracted highly skilled individuals with foreign experience to return at different times. Because firms tend to hire directors located close to their headquarters in the United States and China (Giannetti et al., 2015; Knyazeva et al., 2013), we consider the introduction of the provincial policies exogenous shocks to the supply of potential returnee directors for firms headquartered in those provinces.

Unresolved agency problems can lead corporate managers to choose a level of tax avoidance that is too high or too low relative to what shareholders desire. Returnee directors whose preferences are more closely aligned with shareholders can shift tax avoidance strategies to a more efficient level. Assuming that shareholders have a preference for some, but not over-aggressive tax avoidance strategies, we conjecture that the fraction of returnee directors on the board reduces high tax avoidance and increases tax avoidance that is at too low a level. Consistent with Armstrong et al. (2015), returnee directors who know the net benefits of tax strategies should encourage more tax planning at low levels of tax avoidance and discourage additional tax avoidance when the tax avoidance level is high, and tax avoidance’s marginal costs are higher than its marginal benefits.

To test the hypotheses above, we estimate quantile regressions following Armstrong et al. (2015) to evaluate the relation between tax avoidance and the fraction of returnee directors within a firm. This approach is suitable because the relation may vary in different parts of the tax avoidance distribution. To address the concern that the decision to hire directors with foreign experiences is endogenous and may be influenced by unobserved factors, we construct instrumental variables by using the introduction of the provincial policies to attract returnees. We incorporate the instrumental variable into quantile regression analysis, following the existing literature.

We find that the relation between the fraction of returnee directors on the board and tax avoidance is positive at low levels of tax avoidance and negative at high levels of tax avoidance. It is more likely for a firm to move from the lowest tax avoidance quantiles to higher tax avoidance quantiles as the fraction of returnee directors increases. Similarly, there is a higher likelihood that a firm will move from the highest tax avoidance quantiles to lower tax avoidance quantiles as the fraction of returnee directors increases.

Studying how this relation varies in the cross-section of firms, we examine whether returnee directors’ effects on tax avoidance differ between state-owned enterprises (SOEs) and non-SOEs. We argue that SOEs are more likely to adopt suboptimal tax strategies because SOEs tend to emphasize political objectives rather than economic optimization. We find that the effect of returnee directors on tax avoidance is more positive for SOEs than for non-SOEs when tax avoidance levels are low, and the effect is more negative for SOEs than for non-SOEs when the levels of tax avoidance are high. These results suggest that returnee directors can alleviate economic suboptimality faced by SOEs and help these firms choose a more efficient tax avoidance practice.

We then study the returnee directors’ characteristics that moderate the relation. We find that returnee directors with MBA degrees or directors with a background in accounting or auditing exert a more substantial effect than other returnee directors. These results suggest that returnee directors’ ability and knowledge help identify tax strategies preferred by investors. Furthermore, we find that independent returnee directors can more effectively shift firms’ tax avoidance practices than independent local directors. These results indicate that returnee directors can affect tax avoidance because they better monitor the firms than conventional independent directors do.

Our study contributes to the existing literature on tax avoidance in three ways. First, our study emphasizes board directors’ role in shaping corporate tax avoidance, particularly in fast-changing economies. Our results add to a growing line of research on how board directors affect tax avoidance (Armstrong et al., 2015; Chyz & Gaertner, 2018; Lanis & Richardson, 2011; Minnick & Noga, 2010; Richardson et al., 2013). Second, our study sheds light on how returnees drive the transmission of tax avoidance from one country to another. Our work is related to the prior study about the static cross-nation difference of tax avoidance because of practice and culture norms (Bame-Aldred et al., 2013; Richardson, 2006, 2008; Tsakumis et al., 2007) and extends the studies about the cross-nation transmission of tax avoidance practices (Blouin et al., 2005; Braguinsky & Mityakov, 2015; DeBacker et al., 2015) by focusing on individual migrants. Our findings display a different angle of understanding how individuals are incentivized to engage in corporate tax avoidance from extant literature, which mostly focused on executives (Armstrong et al., 2012; Chyz & Gaertner, 2018; Desai & Dharmapala, 2006; Phillips, 2003; Rego & Wilson, 2012).

The remainder of the study is organized as follows: “The Chinese Environment” section introduces the relevant information about China’s business and regulatory environment. The “Development of Hypotheses” section discusses the development of the hypotheses. The “Sample, Variables, and Identification Strategy” section introduces the sample, variables, and identification strategy used in the study. The section titled “Relation between Returnee Directors and Corporate Tax Avoidance” presents empirical results about the relation between returnee directors and corporate tax avoidance. This section also discusses how the relation between returnee directors and tax avoidance depends on firms’ characteristics and returnee directors. The “Robustness” section evaluates the robustness of the results. The “Concluding Remarks” section concludes our analysis.

The Chinese Environment

Firms listed in China represent the second-largest stock market in the world. These firms have a total market capitalization of US$3.66 trillion by the end of 2012, with 2,494 firms listed on the two stock exchanges in Shanghai and Shenzhen in mainland China. These firms face a substantial number of corporate governance issues, and the interests of minority investors are often compromised. Numerous firms listed in China exhibit low transparency because they often have concentrated ownership and business group affiliations.

A distinct characteristic of Chinese listed firms is the considerable shareholdings retained by the government. Many firms listed in China were spun off from SOEs and then controlled by the government. Shleifer and Vishny (1994) argue that government-owned firms are inefficient because of the imposition of objectives other than profit maximization. Naughton (1995) points out that SOEs tend to emphasize political objectives rather than economic efficiency and are not successful in confronting emerging competition from nonstate enterprises. The complex corporate structure and environment create opportunities for controlling shareholders to take advantage of minority investors. To alleviate some of these concerns, many investors rely on the board of directors.

The board of directors in China functions similarly to the board of directors in the United States. An essential part of the board of directors—the independent directors—was established by the regulators in China by “transplanting” U.S. corporate law and practice (Clarke, 2006). Directors are obliged to review the major plans and policies of their firms. They also oversee the selection, evaluation, and dismissal of top managers.¹ During a board meeting, the directors discuss and vote on proposals in a manner similar to that in the United States. Proposals are approved through voting by directors with simple majority support. Independent directors play an essential role on the board. They have no business relations with the firms they monitor or the controlling shareholders of these firms. Besides, independent directors can serve up to only two terms of 3 years each. They are nominated by large shareholders or the board and formally appointed by shareholders onto the board at a shareholder meeting. Independent directors are typically accomplished individuals who value their reputation inside and outside of the market of board directors (Jiang et al., 2016).

The board is vital for corporate governance in China; however, a shortage of talented directors has been observed. This deficiency is consistent because China has a surplus of unskilled labor and a shortage of individuals experienced in teamwork, foreign languages, international communication, and cultural exposures. This talent shortage can be improved by returning skilled Chinese emigrants who have studied or worked abroad. Studies in international migration flow suggest the growing presence of large numbers of returnees in emerging markets, and China is not an exception. About 2.6 million Chinese have gone abroad to study from 1978 to 2012, and this number has grown in recent years by about 400,000 annually. The majority of them have stayed overseas, whereas 1.1 million have returned (Statistics China, 2013).

Development of Hypotheses

We examine the relation between returnee directors and tax avoidance, focusing on the role of corporate governance in tax avoidance. Following Armstrong et al. (2015), we argue that tax avoidance can be a risky activity for executives to decide. Executives can choose a tax avoidance level that is different from what shareholders desire. Board directors with foreign experience can potentially affect firms’ tax avoidance decisions for at least two reasons. First, returnee directors have been exposed to foreign competition globally and have gained knowledge and ability to shift to a superior management style. Possessing such knowledge and ability, returnee directors can help reduce the deviation of tax avoidance from what shareholders would prefer. Second, directors with foreign experience exhibit the potential to monitor the firm more effectively because their foreign experience renders them relatively independent of the local business environment, governments, and management personnel. Consequently, returnee directors can be better monitors and can identify extreme tax avoidance levels and effectively change tax avoidance.

Using the reasoning above, we conjecture that the relation between tax avoidance and the fraction of returnee directors on the board varies between high and low levels of tax avoidance. Tax planning’s marginal benefits are likely more significant at low levels of tax avoidance than the marginal costs. Directors with foreign experience know the benefits and act as effective monitors; thus, they can encourage tax avoidance, leading to a positive relation between the fraction of returnee directors and tax avoidance at low tax avoidance levels. Tax avoidance’s marginal costs can outweigh the marginal benefits of tax avoidance at high levels of tax avoidance. The potential cost can be regulatory, legal, or political and quickly increase with tax avoidance, whereas the marginal benefits can decrease. In this situation, returnee directors can identify the problem and act as monitors to reduce tax avoidance, leading to a negative relation between the fraction of returnee directors on the board and tax avoidance at high tax avoidance levels.

Hypothesis 1 (H1): The relation between the fraction of returnee directors on the board and tax avoidance is positive at low levels of tax avoidance and negative at high levels of tax avoidance.

If H1 is correct, we expect the effect of returnee directors to vary in the cross-section depending on the characteristics of firms and returnee directors. Examining the cross-sectional variation can provide further evidence related to H1.

We first examine the difference between SOEs and non-SOEs. The government retains considerable shareholdings in Chinese listed firms because many listed firms are carved out from SOEs. The government or its agents often continue to be the controlling shareholders (Gul et al., 2010; Jiang et al., 2016). In our sample, at the end of 2000, the ultimate controlling shareholder of most Chinese firms (75%) was either the central or local government or an SOE. That fraction had decreased to about 40% in 2012. We argue that SOEs could choose tax avoidance strategies that deviate more from what shareholders prefer than non-SOEs because SOEs tend to emphasize objectives other than economic optimization (Naughton, 1995; Shleifer & Vishny, 1994). We expect that returnee directors’ effect on tax avoidance can be more significant for SOEs than for non-SOEs because their knowledge and experience can help firms choose tax avoidance strategies that align with what shareholders prefer.

Hypothesis 2a (H2a): The effect of returnee directors on tax avoidance is more positive at low levels of tax avoidance and more negative at high levels of tax avoidance in SOEs than non-SOEs.

We then consider the educational background and work experience of returnee directors. Specifically, we consider whether a returnee director has an MBA degree. We argue that better business knowledge and skills beyond the corporate tax structure can help returnee directors shift the firm to a better tax avoidance style. An MBA program provides courses on many subjects relevant to a competitive business world, including managerial economics, accounting, business ethics, finance, and many other subjects. With this kind of training, an MBA student can better understand the economic incentives behind tax avoidance. We also consider whether a returnee director has an accounting or auditing background. An experienced accountant or auditor can examine and analyze the firm’s financial health, internal control, and governance and communicate with the firms’ stakeholders about these matters. Such a professional can also identify problems related to tax avoidance. Therefore, we argue that returnee directors with MBA degrees and accounting or auditing backgrounds can better affect tax avoidance practice.

Hypothesis 2b (H2b): The effect of the returnee directors with MBA degrees on tax avoidance is more positive at low levels of tax avoidance and more negative at high levels of tax avoidance than that of the returnee directors without MBA degrees.

Hypothesis 2c (H2c): The effect of the returnee directors with accounting or auditing backgrounds on tax avoidance is more positive at low levels of tax avoidance and more negative at high levels of tax avoidance than that of the returnee directors without accounting or auditing backgrounds.

Independent returnee directors can monitor the firm more effectively because their foreign experience makes them relatively more independent of the local business environment, governments, and management personnel to which independent local directors are closely connected.

Hypothesis 2d (H2d): The effect of independent returnee directors on tax avoidance is more positive at low levels of tax avoidance and more negative at high levels of tax avoidance than that of independent local directors.

Sample, Variables, and Identification Strategy

Sample

We test our hypotheses using a sample of firms listed in China’s A-share market from 2000 to 2012. We collect data on these firms from the China Stock Market and Accounting Research Database (CSMAR). After excluding firms with missing financial information, we obtain a sample of 2,450 firms and 16,197 firm-year observations. We manually collect information on the foreign experience and other demographic characteristics from the biographies of 51,915 directors.

Variables

Tax avoidance variables

Empirical studies have proposed various methods to construct the dependent variable that captures corporate tax avoidance activities (Hanlon & Heitzman, 2010). Following Desai and Dharmapala (2006, 2009), we use the Residual book-tax gap as a primary measure of tax avoidance in this study. The definition and construction of this variable are provided in the subsequent paragraphs.

First, we construct Book–tax gap in accordance with Manzon and Plesko (2002) and Desai and Dharmapala (2006). We calculate the difference between the book income reported by the firm to its shareholders (Y^S) and the income reported by the firm to tax authorities (Y^T), where Y^S is the income from the financial statements of the firm. To estimate Y^T, we divide the firm’s tax expenses (CFTE) by the statutory tax rate (T) faced by the firm², that is, Y^T = CFTE/T. We then obtain Book–tax gap by scaling the difference between Y^S and Y^T with the total assets of the firm. Similar to the method employed by Desai and Dharmapala (2006), we restrict our sample to firm-year observations for which inferred taxable income is positive.³ When the difference between the income reported to shareholders and the income reported to the tax authorities is higher, Book–tax gap is higher, and tax avoidance by the firm becomes more severe. This construction implies that Book–tax gap is related to the difference between statutory tax rate and effective tax rate (ETR). We define the ETR as tax expenses divided by book earnings (T^E= CFTE/Y^S). We define the Book–tax gap as the difference between statutory tax rate and ETR multiplied with income reported to shareholders divided by statutory tax rate. This statement is illustrated in an equation: Y^S−Y^T = Y^S−CFTE/T = Y^S(1 −CFTE/[Y^ST]) = Y.S. (T−T^E)/T. Therefore, Book–tax gap is positively related to the difference between the statutory tax rate and the ETR (T−T^E).

Book–tax gap is widely used in the tax avoidance literature (Desai, 2003; Hanlon, 2005; Lev & Nissim, 2004; Plesko, 2002); however, it may include noise from earnings management (Hanlon & Heitzman, 2010). To alleviate this concern, Desai and Dharmapala (2006, 2009) alternatively measure tax avoidance by using the component of Book–tax gap that is orthogonal to earnings management, which is Residual book-tax gap. Following their method, we use total accruals⁴ to account for the component of the Book–tax gap that is attributable to earnings management. We then estimate the following ordinary least squares regression: Book–tax gap_it = βTotal accruals_it+μ_i+ϵ_it, where μ_i is the average value of firm i’s residuals over the entire sample period from 2000 to 2012. The residual from this regression, ϵ_it, can measure tax avoidance because it represents the component of Book–tax gap that cannot be explained by variations in earnings management. Specifically, we define Residual book-tax gap as Residual book-tax gap_it = μ_i+ϵ_it.

We industry-size adjust the measures of tax avoidance, Residual book-tax gap. Specifically, we sort firms into size deciles each year using total assets, calculate the mean of the tax avoidance variable within each industry and size decile combination each year, and subtract the mean from the variable before estimating the regressions. The industry-size adjusted variables can better control for the industry- and size-specific shocks. For example, if an industry or a size group exhibits an increasing trend in tax avoidance, the industry-size-year mean captures that trend. We remove the trend from the variation in the tax avoidance variables by using industry-size-year mean adjusted tax avoidance. Following Giannetti et al. (2015), we use the China Securities Regulatory Commission’s official industry classification to define industries.

Returnee director variables

Our primary independent variable is Returnee directors. This variable is the number of Chinese directors with foreign experience divided by the total number of directors. An individual gains foreign experience if that person has studied or worked outside mainland China.⁵ We obtain information regarding the foreign experience of the directors by examining their biographies from CSMAR. A total of 51,915 unique directors and 179,320 director-firm-year observations are listed, and 9.07% of all directors have foreign experience. Most returnee directors have earned their overseas experience in Hong Kong, followed by those who have obtained experience in the United States, United Kingdom, Japan, Canada, Australia, Singapore, Germany, France, and Macau. The detailed geographic distribution showing the returnee director’s destinations is presented in IA Table 2 in the Internet Appendix.

We subsequently divide Returnee directors into two mutually exclusive variables according to two characteristics of the director. First, we focus on whether the returnee director obtains an MBA degree and then divide Returnee directors into Returnee directors with MBA degrees and Returnee directors without MBA degrees. The former represents the number of returnee directors with an MBA degree, divided by the number of board members; the latter is the number of returnee directors without an MBA degree, divided by the number of board members. Second, we consider returnee directors with a background in accounting or auditing. Explicitly, we calculate Returnee directors with A-A background as the number of returnee directors with a background in accounting or auditing, divided by the number of board directors. We then calculate Returnee directors without A-A background as the number of returnee directors without such a background, divided by the number of board directors. A director is considered to have a background in accounting or auditing if that individual has studied accounting or auditing or worked in related areas.⁶

Furthermore, we measure the returnee directors’ relative independence by creating three variables: Independent returnee directors, Independent local directors, and Nonindependent returnee directors. Independent returnee directors represents the number of returnee directors who are not employees of the firm, divided by the number of board directors. Independent local directors is the number of directors without foreign experience who are not employees of the firm, divided by the number of board directors. Nonindependent returnee directors is the number of returnee directors who are employees of the firm, divided by the number of board directors.

Control variables

According to previous studies, we construct several firm-level variables that may influence corporate tax avoidance decisions (e.g., Dyreng et al., 2010; Frank et al., 2009). All these measures are obtained from the CSMAR database. First, we include a measure of foreign ownership, which is defined as the shares owned by all foreign legal persons divided by the firm’s total shares (Foreign ownership). The share fraction of the largest shareholder (Block) and a dummy variable that is one of their largest ultimate shareholders are marked in the CSMAR database with “state-owned” or “state-controlled” (State). These three variables capture the ownership structure of the firm. Second, we use the natural logarithm of the firm’s total assets to measure the firm size (Firm size) and use the firm’s operating income divided by the total assets (return-on-assets [ROA]) to measure profitability. These variables can capture the effect that smaller and more profitable firms are more likely to engage in tax avoidance. Third, we include as controls the total liabilities divided by total assets (Leverage), the intangible assets divided by total assets (Intangible), and the sum of cash and bank balance divided by total assets (Cash). We include the last variable because tax avoidance can depend on the firm’s availability of cash resources.

We also include several board-level variables related to the firm’s tax avoidance decisions. Director age is the board-level average of the difference between the current year and the director’s birth year. Director tenure is defined as the board-level average of one plus the difference between the current year and the year when the individual joined the board. Female director is the fraction of directors who are female. Nonindependent director is the fraction of directors who receive a salary as an employee of the firm. Foreign director is the fraction of directors on the board who are foreigners.

We winsorize all the above variables at the first and 99th percentiles, except for discrete indicator variables. To better examine the impact of returnee directors on the contemporaneous tax avoidance, we follow Armstrong et al. (2015) to lag all returnee director variables and control variables related to the board directors by one year in the following analyses.

Table 1 lists summary statistics for the key variables in our analysis, and the table presents descriptive statistics separately for firms with returnee directors and those without returnee directors. The average Residual book-tax gap is 0.0001 for the firm-year observations with returnee directors and −0.0005 for the firm-year observations without returnee directors. The difference in mean Residual book-tax gap is 0.0006 and statistically significant. The difference in average Residual book-tax gap between these two subsamples does not mean that returnee directors prompt firms to avoid more taxes for two reasons. First, we argue that the relation between Returnee directors and tax avoidance is different in different parts of tax avoidance distribution. Thus, we use quantile regressions to allow the relation to vary across different parts of the tax avoidance distribution. Second, firms in these two subsamples can differ in other firm characteristics that can drive tax avoidance. Therefore, we control for a series of other variables and use an instrumental variable approach. Table 1 also shows that Returnee directors has a mean of 19.86% and a standard deviation of 12.46% among the 7,974 firm-year observations with returnee directors and is equal to zero for the remaining 6,108 observations without returnee directors. Returnee directors have a mean of 11.48% in the full sample, which suggests that one of 10 directors is a returnee director.

Table 1.

Summary Statistics.

Group	With returnee directors			Without returnee directors			Difference of mean
Variable	M	SD	No. of obs	M	SD	No. of obs	Difference of mean
Residual book-tax gap	0.0001	0.0049	7,681	−0.0005	0.0047	5,737	0.0006***
Returnee directors	0.1986	0.1246	7,974	0	0	6,108	0.1986***
Number of returnee directors	1.9437	1.3006	7,974	0	0	6,108	1.9437***
Director age	48.9497	3.9184	7,971	48.5706	3.9220	6,105	0.3791***
Director tenure	2.9026	0.4749	7,974	2.8914	0.4371	6,108	0.0112
Female director	0.1009	0.1029	7,974	0.1095	0.1085	6,108	−0.0087***
Nonindependent director	0.3694	0.1956	7,969	0.4035	0.2168	6,101	−0.0341***
Foreign director	0.0077	0.0330	7,974	0.0004	0.0087	6,108	0.0072***
Foreign ownership	0.0153	0.0630	7,974	0.0038	0.0297	6,108	0.0115***
Block	0.3796	0.1582	7,974	0.3932	0.1645	6,108	−0.0136***
State	0.4105	0.4919	7,974	0.5164	0.4998	6,108	−0.1059***
Firm size (natural logarithm of numbers in RMB)	21.7435	1.2415	7,974	21.4254	1.0231	6,108	0.3181***
ROA	0.0544	0.0547	7,974	0.0469	0.0549	6,108	0.0075***
Leverage	0.4768	0.2309	7,974	0.4801	0.2393	6,108	−0.0033
Intangible (divided by total assets)	0.0414	0.0530	7,974	0.0424	0.0545	6,108	−0.001
Cash (divided by total assets)	0.1747	0.1338	7,699	0.1605	0.1167	5,977	0.0142***
Returnee directors with MBA degrees	0.0182	0.0482	7,974	0	0	6,108	0.0182***
Returnee directors without MBA degrees	0.1804	0.1154	7,974	0	0	6,108	0.1804***
Returnee directors with A-A background	0.0316	0.0549	7,974	0	0	6,108	0.0316***
Returnee directors without A-A background	0.1670	0.1160	7,974	0	0	6,108	0.1670***
Independent returnee directors	0.1482	0.1046	7974	0	0	6,108	0.1482***
Independent local directors	0.4821	0.1932	7,969	0.5962	0.2174	6,101	−0.1141***
Nonindependent returnee directors	0.0504	0.0849	7,974	0	0	6,108	0.0504***
SOE	0.5712	0.4949	7,947	0.6684	0.4708	6,094	−0.0972***

Note. This table reports the descriptive statistics for firms with and without returnee directors and difference in mean between the two groups. The detailed definitions of variables are in IA Table 1 in the Internet Appendix. ROA = Return-on-assets; SOE = State-owned enterprise; RMB = Renminbi.

Identification Strategy

In an ideal experiment, directors with foreign experience would be randomly assigned to firms. Researchers observe how firms’ tax avoidance behaviors with returnee directors change relative to those without such directors. However, the decision to hire directors with foreign experience is endogenous and may be influenced by unobserved factors correlated with tax avoidance behaviors. In particular, firms with particular characteristics or firms that encounter particular challenges or opportunities might select or attract board members with foreign experience. These firms can engage in tax avoidance activities in specific ways. As a result, returnee directors’ existence and tax avoidance can be driven by common unobserved factors. Thus, merely controlling for time-invariant industry and province characteristics and firm-level ownership, size, and board structure variables may not identify the causal impact of Returnee directors on tax avoidance. To address the potential endogeneity in tax avoidance decisions, we use an instrumental variable approach.

Following Giannetti et al. (2015), we construct instrumental variables by using the introduction of provincial policies for attracting returnees. Starting in the late 1990s, Chinese provincial governments adopted policies to attract highly skilled individuals with foreign experience, and they did so at different times (Zweig, 2006). The policy’s main objective was to increase the quality of academic and industrial research and foster entrepreneurial activity and the entry of new businesses. Directed only to the most distinguished Chinese expatriates, the policies included tax breaks, subsidized housing, tax-free imports of automobiles and computers, schooling for the children of the returnees, local grants and awards, medical benefits, jobs for spouses, and long-term residence permits. These provincial policies are exogenous shocks in the supply of potential directors with foreign experience for firms headquartered in those provinces because firms tend to hire directors located close to their headquarters in the United States and China (Giannetti et al., 2015; Knyazeva et al., 2013). Therefore, this instrumental variable can help us address the endogeneity issue.⁷

The instrumental variables include Provincial policy and the interaction between Provincial policy and the ex ante ownership characteristic variables of the firm (State, Foreign ownership, and Block) in 2000, the beginning of our sample period. Provincial policy is a variable equal to one in years following the policy’s implementation; otherwise, this variable is zero. If a firm enters our sample after 2000, the firm ownership characteristics during the year the firm entered the sample are used to construct the interaction variables. We add the interaction terms to increase the precision of the first stage of the two-stage least square estimation in the instrumental variable approach because the effect of Provincial policy on Returnee directors can depend on the ex ante ownership characteristic variables.⁸ These instrumental variables capture changes in the supply of directors with the foreign experience driven by the policies because the labor market for board directors is local.⁹ The policies to attract highly skilled returnee migrants thus lead to arguably exogenous increases in the supply of potential directors with foreign experience in different provinces at different times. Hence, we expect these variables to be reliable predictors of an increase in Returnee directors.¹⁰

We incorporate the instrumental variables into quantile regression estimation. The previous literature considers using instrumental variables with quantile regressions (Abadie et al., 2002; Chernozhukov & Hansen, 2005; Chesher, 2003; Lee, 2007; Frölich & Melly, 2013; Kaplan & Sun, 2017; Powell, 2013). We follow the control function approach by Lee (2007).

The Relation Between Returnee Directors and Corporate Tax Avoidance

We examine H1, which states that the relation between the director’s foreign experience and tax avoidance is positive at low tax avoidance levels and negative at high levels of tax avoidance. To evaluate this hypothesis, we use quantile regression of industry-size-year mean adjusted Residual book-tax gap on Returnee directors and control variables.

\begin{matrix} Tax avoidanc e_{i, t} = β_{0} + β_{1} Returnee director s_{i, t - 1} + β_{2} Foreign ownershi p_{i, t} \\ + β_{3} Bloc k_{i, t} + β_{4} Stat e_{i, t} + β_{5} Firm siz e_{i, t} + β_{6} Leverag e_{i, t} + β_{7} RO A_{i, t} \\ + β_{8} Intangibl e_{i, t} + β_{9} Cas h_{i, t} + β_{10} Female directo r_{i, t - 1} \\ + β_{11} Director tenur e_{i, t - 1} + β_{12} Director ag e_{i, t - 1} + β_{13} Foreign directo r_{i, t - 1} \\ + β_{14} Nonindependent directo r_{i, t - 1} + ε_{i, t} . \end{matrix}

We follow Armstrong et al. (2015) to lag Returnee directors and control variables related to board directors by one year. The other control variables are contemporary with the dependent variable. Considering that the supply and employment of the returnee directors can be endogenously related to firms’ characteristics where they work as directors, we present results of the second stage of quantile regressions with instrumental variables in Panel A of Table 2. The 10th percentile coefficient is 0.0860 with a t stat of 3.14, and that at the 90th percentile is −0.1230 with a t stat of −5.45. The F test statistics show that coefficients of Returnee directors are significantly different between the 0.1 and 0.9 quantiles, between 0.2 and 0.8 quantiles, between 0.3 and 0.7 quantiles, or between 0.4 and 0.6 quantiles. These results in Panel A suggest that after controlling for the potential endogeneity, returnee directors will increase tax avoidance when its level is relatively low and decrease tax avoidance when its level is relatively high.¹¹ The economic significance of the coefficients is large in the left and right tails. A one standard deviation increase in Returnee directors is associated with an increase in the 10th percentile of Residual book-tax gap equal to 1.18 standard deviations of Residual book-tax gap and a decrease in the 90th percentile of Residual book-tax gap equal to 1.69 standard deviations of Residual book-tax gap. These results suggest that firms with more returnee directors tend to increase tax avoidance when they are too conservative in the book-tax gap relative to those same-industry and similar-size firms. These firms decrease tax avoidance when the firms are potentially too aggressive relative to those comparable firms. These results are consistent with H1.

We discuss the coefficients of some control variables, which are listed in Panel A of Table 2. We find that smaller firms consistently engage in more tax avoidance across the tax avoidance distribution. This result is similar to the findings by Zimmerman (1983), Slemrod (2004), and Cai and Liu (2009). We also find that the association between the fraction of nonindependent directors (Nonindependent director) and tax avoidance is significantly negative in the left tail of the tax avoidance distribution and positive in the right tail. This result suggests that when a firm has a lower fraction of independent directors, it further reduces tax avoidance even at low levels of tax avoidance and increases tax avoidance even at high levels of tax avoidance. This result is consistent with the findings by Armstrong et al. (2015). Furthermore, the coefficients of State in Panel A indicate that enterprises owned by the state avoid more tax than private firms in most quantiles, which is consistent with the findings by Tang (2016).

Table 2.

Returnee Director and Corporate Tax Avoidance.

Panel A: Quantile Regressions With Instrumental Variables.
Quantile	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
IDV	DV = Residual book-tax gap with industry-sized adjustment, n = 12,966
Returnee directors	0.0860***	0.0485**	0.0061	−0.0084	−0.0273**	−0.0519***	−0.0783***	−0.0963***	−0.1230***
	(3.14)	(2.31)	(0.34)	(−0.65)	(−2.28)	(−2.76)	(−4.65)	(−3.92)	(−5.45)
Foreign ownership	−0.0160	−0.0052	0.0055	0.0181***	0.0250***	0.0346***	0.0420***	0.0437***	0.0485***
	(−1.18)	(−0.49)	(0.75)	(3.87)	(5.20)	(5.63)	(6.54)	(5.47)	(6.28)
Block	−0.0003	0.0005	0.0009	0.0008	0.0005	0.0000	−0.0001	−0.0034**	−0.0054*
	(−0.07)	(0.19)	(0.41)	(0.45)	(0.31)	(0.02)	(−0.04)	(−2.05)	(−1.69)
State	0.0028**	0.0017**	0.0012	0.0013*	0.0012	0.0013	0.0003	0.0003	0.0005
	(2.33)	(1.97)	(1.28)	(1.85)	(1.55)	(1.63)	(0.38)	(0.29)	(0.46)
Firm size	−0.0121***	−0.0086***	−0.0060***	−0.0050***	−0.0044***	−0.0039***	−0.0035***	−0.0037***	−0.0031***
	(−12.35)	(−13.40)	(−12.07)	(−10.64)	(−13.22)	(−6.96)	(−6.20)	(−5.05)	(−4.36)
Leverage	0.0296***	0.0188***	0.0156***	0.0139***	0.0131***	0.0145***	0.0142***	0.0153***	0.0115***
	(5.82)	(7.33)	(7.86)	(7.07)	(7.82)	(10.37)	(8.53)	(8.26)	(5.06)
ROA	0.7256***	0.7537***	0.7681***	0.7723***	0.7692***	0.7724***	0.7606***	0.7449***	0.7338***
	(42.32)	(82.80)	(81.78)	(118.49)	(93.11)	(106.07)	(106.12)	(111.63)	(82.15)
Intangible	0.0101	0.0091	0.0047	−0.0003	−0.0060	−0.0073*	−0.0132***	−0.0178**	−0.0259***
	(1.01)	(1.23)	(0.98)	(−0.07)	(−1.41)	(−1.65)	(−2.82)	(−2.08)	(−2.85)
Cash	−0.0033	−0.0068*	−0.0004	0.0027	0.0049**	0.0083***	0.0119***	0.0185***	0.0198***
	(−0.77)	(−1.94)	(−0.15)	(1.37)	(2.06)	(3.68)	(3.97)	(5.87)	(6.64)
Female director	−0.0015	−0.0065	−0.0069***	−0.0061***	−0.0032	−0.0040*	−0.0050*	−0.0057*	−0.0065*
	(−0.24)	(−1.64)	(−2.68)	(−2.69)	(−1.49)	(−1.76)	(−1.86)	(−1.90)	(−1.68)
Director tenure	0.0000	−0.0012**	−0.0007	−0.0007	−0.0003	−0.0002	−0.0004	−0.0001	0.0000
	(0.02)	(−1.99)	(−1.44)	(−1.26)	(−0.52)	(−0.42)	(−0.50)	(−0.13)	(0.02)
Director age	−0.0001	0.0000	−0.0001*	−0.0001	−0.0001	−0.0000	−0.0000	−0.0000	−0.0002**
	(−0.66)	(0.33)	(−1.65)	(−0.94)	(−0.84)	(−0.02)	(−0.42)	(−0.16)	(−2.31)
Foreign director	−0.0636**	−0.0222	0.0162	0.0255	0.0413***	0.0709***	0.0919***	0.1177***	0.1443***
	(−2.00)	(−0.80)	(0.66)	(1.52)	(2.67)	(3.35)	(4.14)	(4.30)	(4.42)
Nonindependent director	−0.0082***	−0.0048**	−0.0035**	−0.0013	−0.0007	0.0001	0.0022	0.0032	0.0015
	(−2.95)	(−2.55)	(−1.99)	(−1.04)	(−0.46)	(0.09)	(1.17)	(1.38)	(0.68)
Intercept	0.2537***	0.1696***	0.1126***	0.0837***	0.0678***	0.0485***	0.0381***	0.0350***	0.0278**
	(16.16)	(15.28)	(12.25)	(10.31)	(9.24)	(4.36)	(3.43)	(2.64)	(2.34)
Province fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pseudo R²	.3827	.4052	.4052	.3972	.3850	.3700	.3510	.3278	.2971
Test hypothesis							t stat	p value
(Q 0.10) Returnee directors = (Q 0.90) Returnee directors							32.42***	.0000
(Q 0.20) Returnee directors = (Q 0.80) Returnee directors							40.99***	.0000
(Q 0.30) Returnee directors = (Q 0.70) Returnee directors							21.12***	.0000
(Q 0.40) Returnee directors = (Q 0.60) Returnee directors							9.44***	.0021

Panel B: Interquantile Movement.
	2SLS DV = Quantile of residual book-tax gap with industry-sized adjustment, n = 10,906
Independent variable	(1)	(2)
Lagged quantile	0.1170***	0.2321***
	(9.22)	(7.06)
Returnee directors	−4.4201	2.4992
	(−1.53)	(0.77)
Lagged quantile × returnee directors		−1.0499***
		(−3.81)
Controls	Yes	Yes
Firm fixed effect	Yes	Yes
# of firms	1,839	1,839
Adjusted R²	.2942	.2794

Note. Panel A considers the endogeneity and reports the second stage of quantile regressions with instrumental variables. The regression equation is as follows: $\begin{matrix} Tax avoidanc e_{i, t} = β_{0} + β_{1} Returnee director s_{i, t - 1} + β_{2} Foreign owne r shi p_{i, t} \\ + β_{3} Bloc k_{i, t} + β_{4} Stat e_{i, t} + β_{5} Firm siz e_{i, t} + β_{6} Leverag e_{i, t} + β_{7} RO A_{i, t} \\ + β_{8} Intangibl e_{i, t} + β_{9} Cas h_{i, t} + β_{10} Female directo r_{i, t - 1} \\ + β_{11} Director tenur e_{i, t - 1} + β_{12} Director ag e_{i, t - 1} + β_{13} Foreign directo r_{i, t - 1} \\ + β_{14} Nonindependent directo r_{i, t - 1} + ε_{i, t} . \end{matrix}$

Panel B reports the second stage of 2SLS regression of the current quantile of Residual book-tax gap with industry-sized adjustment on the tax avoidance quantile in the previous year, returnee directors, and the interaction between them. For Panel A, two-sided t stats and p values are reported for the tests of coefficient differences between quantiles. 2SLS = two-stage least squares; DV = dependent variable; IDV = independent variable; ROA = return-on-assets.

The significance signs *, **, and *** mean p < .1, <.05, and <.01, respectively.

To provide more support for H1, we study the interquantile movement to paint the whole picture about the impact of Returnee directors on the firm’s choice of tax avoidance level. First, we sort firms within each industry into deciles by tax avoidance for the current year and the previous year. Second, we regress the current tax avoidance decile on the previous tax avoidance decile and the interaction between the last tax avoidance quantile and Returnee directors. The results of these regressions are in Panel B of Table 2. Column (1) shows that the coefficient on the last decile of tax avoidance is positive and less than one, after controlling for those firm- and board-level variables. This coefficient means that the level of firms’ tax avoidance tends to be mean-reverting. This result implies that firms with relatively high tax avoidance ranking in the previous year tend to have lower tax avoidance ranking in the following year because the positive distance between the tax avoidance last year and its long-run average is multiplied with a coefficient less than one and will be shorter. Similarly, firms with relatively low tax avoidance ranking tend to have higher tax avoidance quantiles in the next year.

Furthermore, in Column (2), the interaction term’s coefficient is negative and significant, which means that the previous tax avoidance quantile’s effective coefficient is smaller when the fraction of returnee directors in the last year is relatively high. This result implies that the mean-reverting effect of tax avoidance quantiles is more severe when Returnee directors is relatively high. Therefore, there is a higher likelihood a firm will move from high tax avoidance quantiles to low tax avoidance quantiles when Returnee directors is relatively high. Similarly, it is more likely that a firm will move from low tax avoidance quantiles to high tax avoidance quantiles when Returnee directors is relatively high. The results suggest that returnee directors “move” tax avoidance of firms based on the current level of tax avoidance, which is consistent with H1.

After documenting the relation between Returnee directors and tax avoidance, we further study how this relation varies in the cross-section of firms and returnee directors. First, in Table 3 Panel A, we test whether returnee directors’ effects on tax avoidance quantiles differ between SOEs and non-SOEs. We construct a dummy variable SOE, which equals to one if the ultimate controlling shareholder of the firm is marked in the CSMAR database as “central government,”“local government,” or “SOE.” Second, we estimate IV quantile regressions of tax avoidance on Returnee directors, SOE, and the interaction between Returnee directors and SOE. The coefficients of Returnee directors show a robust trend, as shown in Table 3 Panel A. The interaction’s coefficients show a similar trend: It is significantly positive at the 10th percentile (0.0255) and significantly negative at the 90th percentile (−0.0174). The difference in the coefficients of the interaction between the 10th and 90th percentiles is statistically significant, with a p value of .0000. Returnee directors’ effects on tax avoidance quantiles are stronger for SOEs than for non-SOEs, suggesting that returnee directors can help these SOEs choose the more efficient tax avoidance practice. These results show support for H2a.

Table 3.

Evidence in the Cross-Sections of Firms and Returnee Directors.

Panel A: Effects of Returnee Directors in SOEs and Non-SOEs.
Quantile	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
IDV	DV = Residual book-tax gap with industry-sized adjustment, n = 12,935
Returnee directors	0.0857**	0.0357*	−0.0031	−0.0151	−0.0384***	−0.0778***	−0.1025***	−0.1164***	−0.1459***
	(2.42)	(1.85)	(−0.24)	(−1.18)	(−3.59)	(−4.54)	(−7.60)	(−5.74)	(−7.09)
Returnee directors × SOE	0.0255***	0.0175***	0.0132***	0.0050	0.0017	0.0012	−0.0028	−0.0055	−0.0174***
	(4.35)	(3.86)	(3.03)	(1.22)	(0.44)	(0.28)	(−0.69)	(−0.99)	(−2.69)
SOE	−0.0011	−0.0014	−0.0015*	−0.0006	−0.0014	−0.0021*	−0.0028***	−0.0026***	−0.0010
	(−0.51)	(−1.03)	(−1.92)	(−0.58)	(−1.62)	(−1.94)	(−3.00)	(−2.59)	(−1.06)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pseudo R²	.3866	.4080	.4074	.3988	.3864	.3711	.3525	.3295	.2996
Test hypothesis							t stat	p value
(Q 0.10) Returnee directors = (Q 0.90) Returnee directors							27.00***	.0000
(Q 0.20) Returnee directors = (Q 0.80) Returnee directors							64.15***	.0000
(Q 0.30) Returnee directors = (Q 0.70) Returnee directors							54.36***	.0000
(Q 0.40) Returnee directors = (Q 0.60) Returnee directors							26.03***	.0000
(Q 0.10) Returnee directors × SOE = (Q 0.90) Returnee directors × SOE							45.11***	.0000
(Q 0.20) Returnee directors × SOE = (Q 0.80) Returnee directors × SOE							13.69***	.0002
(Q 0.30) Returnee directors × SOE = (Q 0.70) Returnee directors × SOE							14.55***	.0001
(Q 0.40) Returnee directors × SOE = (Q 0.60) Returnee directors × SOE							1.64	.2010

Panel B: The Effects of Returnee Directors With and Without MBA Degrees.
Quantile	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
IDV	DV = Residual book-tax gap with industry-sized adjustment, n = 12,966
Returnee directors with MBA degrees	1.3798**	0.1092	−0.5438	−1.1570***	−1.2988***	−1.9945***	−2.5511***	−3.1136***	−3.6965***
	(1.97)	(0.22)	(−1.57)	(−3.51)	(−3.82)	(−5.60)	(−9.19)	(−10.98)	(−8.31)
Returnee directors without MBA degrees	−0.0852	0.0408	0.0774**	0.1405***	0.1351***	0.2010***	0.2416***	0.3096***	0.3403***
	(−1.16)	(0.62)	(1.99)	(3.39)	(3.63)	(4.44)	(6.53)	(7.33)	(6.19)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Returnee directors with MBA degrees = Returnee directors without MBA degrees	3.65*	0.01	2.66	12.35***	14.54***	30.35***	80.93***	114.15***	65.82***
Pseudo R²	.3830	.4052	.4053	.3975	.3855	.3710	.3530	.3305	.3012
Test hypothesis							t stat	p value
(Q 0.10) Returnee directors with MBA degrees = (Q 0.90) Returnee directors with MBA degrees							34.76***	.0000
(Q 0.20) Returnee directors with MBA degrees = (Q 0.80) Returnee directors with MBA degrees							34.60***	.0000
(Q 0.30) Returnee directors with MBA degrees = (Q 0.70) Returnee directors with MBA degrees							26.71***	.0000
(Q 0.40) Returnee directors with MBA degrees = (Q 0.60) Returnee directors with MBA degrees							12.46***	.0004
(Q 0.10) Returnee directors without MBA degrees = (Q 0.90) Returnee directors without MBA degrees							23.27***	.0000
(Q 0.20) Returnee directors without MBA degrees = (Q 0.80) Returnee directors without MBA degrees							14.01***	.0002
(Q 0.30) Returnee directors without MBA degrees = (Q 0.70) Returnee directors without MBA degrees							10.62***	.0011
(Q 0.40) Returnee directors without MBA degrees = (Q 0.60) Returnee directors without MBA degrees							2.68	.1016

Panel C: The Effects of Returnee Directors With and Without the Background of Accounting or Auditing.
Quantile	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
IDV	DV = Residual book-tax gap with industry-sized adjustment, n = 12,966
Returnee directors with A-A background	0.4423***	−0.0192	−0.1172	−0.2488***	−0.3444***	−0.5131***	−0.6929***	−0.8411***	−0.9657***
	(3.73)	(−0.14)	(−1.30)	(−2.83)	(−4.79)	(−5.41)	(−7.78)	(−11.25)	(−9.46)
Returnee directors without A-A background	0.0029	0.0558*	0.0295	0.0434**	0.0340***	0.0573***	0.0672**	0.0781***	0.0597*
	(0.07)	(1.84)	(1.16)	(2.38)	(2.61)	(2.87)	(2.56)	(3.13)	(1.90)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Returnee directors with A-A background = Returnee directors without A-A background	10.20***	0.24	1.84	8.18***	24.07***	28.13***	48.09***	98.09***	65.57***
Pseudo R²	.3830	.4053	.4053	.3975	.3855	.3710	.3528	.3305	.3010
Test hypothesis							t stat	p value
(Q 0.10) Returnee directors with A-A background = (Q 0.90) Returnee directors with A-A background							94.42***	.0000
(Q 0.20) Returnee directors with A-A background = (Q 0.80) Returnee directors with A-A background							26.21***	.0000
(Q 0.30) Returnee directors with A-A background = (Q 0.70) Returnee directors with A-A background							18.12***	.0000
(Q 0.40) Returnee directors with A-A background = (Q 0.60) Returnee directors with A-A background							6.61**	.0102
(Q 0.10) Returnee directors without A-A background = (Q 0.90) Returnee directors without A-A background							1.15	.2828
(Q 0.20) Returnee directors without A-A background = (Q 0.80) Returnee directors without A-A background							0.35	.5553
(Q 0.30) Returnee directors without A-A background = (Q 0.70) Returnee directors without A-A background							0.79	.3726
(Q 0.40) Returnee directors without A-A background = (Q 0.60) Returnee directors without A-A background							0.23	.6350

Panel D: Role as Independent Returnee Directors.
Quantile	0.10	0.20	0.30	0.40	0.50	0.60	0.70	0.80	0.90
IDV	DV = Residual book-tax gap with industry-sized adjustment, n = 12,966
Independent returnee directors	0.1846***	0.0170	0.0005	−0.0377**	−0.0679***	−0.1042***	−0.1585***	−0.1975***	−0.2385***
	(3.65)	(0.48)	(0.02)	(−2.05)	(−3.50)	(−5.92)	(−9.25)	(−8.27)	(−8.98)
Independent local directors	0.0335***	0.0147***	0.0045	−0.0019	−0.0084***	−0.0160***	−0.0253***	−0.0303***	−0.0396***
	(3.97)	(2.67)	(1.29)	(−0.54)	(−2.72)	(−4.05)	(−6.35)	(−6.45)	(−11.55)
Nonindependent returnee directors	0.0184	0.1056**	0.0211	0.0251	0.0043	−0.0098	−0.0167	0.0104	−0.0303
	(0.30)	(2.51)	(0.64)	(1.16)	(0.16)	(−0.25)	(−0.38)	(0.25)	(−0.85)
Control variables	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Province fixed effect	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Independent returnee directors = Independent local directors	11.99***	0.01	0.03	5.16**	12.14***	33.94***	78.24***	64.25***	66.89***
Independent returnee directors = Nonindependent returnee directors	3.82*	2.10	0.18	5.07**	3.97**	5.11**	8.87***	17.28***	14.31***
Pseudo R²	.3833	.4054	.4053	.3973	.3852	.3704	.3520	.3293	.2992
Test hypothesis							t stat	p value
(Q 0.10) Independent returnee directors = (Q 0.90) Independent returnee directors							84.06***	.0000
(Q 0.20) Independent returnee directors = (Q 0.80) Independent returnee directors							43.80***	.0000
(Q 0.30) Independent returnee directors = (Q 0.70) Independent returnee directors							45.43***	.0000
(Q 0.40) Independent returnee directors = (Q 0.60) Independent returnee directors							19.05***	.0000
(Q 0.10) Independent local director = (Q 0.90) Independent local director							74.13***	.0000
(Q 0.20) Independent local director = (Q 0.80) Independent local director							58.24***	.0000
(Q 0.30) Independent local director = (Q 0.70) Independent local director							48.49***	.0000
(Q 0.40) Independent local director = (Q 0.60) Independent local director							20.11***	.0000
(Q 0.10) Nonindependent returnee directors = (Q 0.90) Nonindependent returnee directors							0.47	.4949
(Q 0.20) Nonindependent returnee directors = (Q 0.80) Nonindependent returnee directors							5.02**	.0250
(Q 0.30) Nonindependent returnee directors = (Q 0.70) Nonindependent returnee directors							0.87	.3523
(Q 0.40) Nonindependent returnee directors = (Q 0.60) Nonindependent returnee directors							1.26	.2624

Note. Panel A presents IV quantile regression examining whether the impact of Returnee directors on Residual book-tax gap with industry-sized adjustment is different in SOEs and private firms. Panel B reports IV quantile regression estimates of Residual book-tax gap with industry-sized adjustment on Returnee directors with MBA degrees. Panel C shows IV quantile regression estimates of Residual book-tax gap with industry-sized adjustment on Returnee directors with background of accounting or auditing. Panel D reports IV quantile regression estimates of Residual book-tax gap with industry-sized adjustment on Independent returnee directors, Independent local directors, and Nonindependent returnee directors; all quantile regressions include various control variables, province fixed effects, and a constant parameter, the coefficients of which are not tabulated. Two-sided t stats and p values are reported for the tests of coefficient differences between quintiles.

The significance signs *, **, and *** mean p < .1, <.05, and <.01, respectively. DV = dependent variable; IDV = independent variable; SOEs = state-owned enterprises.

We then consider three characteristics of returnee directors. The first measure is whether the returnee director has an MBA degree. We examine whether returnee directors with an MBA degree drive the effect of returnee directors on tax avoidance. Returnee directors with MBA degrees is the number of returnee directors with an MBA degree divided by the total number of directors. Returnee directors without MBA degrees is the number of returnee directors without such a degree, divided by the total number of directors. We then replace the original variable, Returnee directors, with Returnee directors with MBA degrees and Returnee directors without MBA degrees in the quantile regressions with instrumental variables. The results presented in Panel B of Table 3 show that the relation between Returnee directors and tax avoidance is driven by directors with MBA degrees.

The effects of Returnee directors with MBA degrees are positive when the levels of tax avoidance are low and negative when the levels of tax avoidance are high. The coefficient of Returnee directors with MBA degrees at the 10th percentile is 1.3798, and the coefficient at the 90th percentile is −3.6965. These effects are large in magnitude and robust statistically. The F test statistics show that the coefficients of Returnee directors with MBA degrees at the 0.1 and 0.9 quantiles are significantly different. Similarly, the differences in coefficients are significant between 0.2 and 0.8 quantiles, between 0.3 and 0.7 quantiles, or between 0.4 and 0.6 quantiles. By contrast, the effects of Returnee directors without MBA degrees are positive and significant for most quantiles. However, the magnitudes of the coefficients of Returnee directors without MBA degrees are much smaller than those of Returnee directors with MBA degrees, meaning that the effects of Returnee directors without MBA degrees are less important economically. The evidence that Returnee directors without MBA degrees is positively related to tax avoidance at most quantiles of tax avoidance could mean that returnee directors without proper business training choose to increase tax avoidance regardless of the tax avoidance level. Overall, these results show that returnee directors with an MBA degree drive the relation between returnee directors’ foreign experience and tax avoidance. In contrast, those without MBA are less or even reverse influential. These results support H2b.

The second characteristic of returnee directors is whether the returnee director has studied accounting or auditing or has worked in related areas. We define Returnee directors with A-A background as the number of returnee directors with a background in accounting or auditing, divided by the total number of directors. Returnee directors without A-A background represents the number of returnee directors without such background, divided by the total number of directors. We then replace Returnee directors in the quantile regressions with these two variables. These two variables’ coefficients can show which returnee directors drive the relation between Returnee directors and tax avoidance. These results are presented in Panel C of Table 3. The coefficient of Returnee directors with A-A background at the 10th percentile is 0.4423, and that at the 90th percentile is −0.9657. The difference between these two coefficients is statistically significant, with a p value of .0000. The coefficient of Returnee directors without A-A background is 0.0029 at the 10th percentile and 0.0597 at the 90th percentile. The difference between these coefficients is not statistically significant, with a p value of .2828. The results above demonstrate that the coefficients of Returnee directors with A-A background are of larger economic magnitudes than those of Returnee directors without A-A background. Moreover, the coefficients of Returnee directors with A-A background are positive at low quantiles of tax avoidance and negative at high quantiles of tax avoidance, similar to the pattern observed in Returnee directors. By contrast, the coefficients of Returnee directors without A-A background exhibit no such a pattern. These results are consistent with H2c.

The third characteristic of returnee directors is whether the returnee director is an independent director. We construct Independent returnee directors and Nonindependent returnee directors. Nonindependent returnee directors represents the number of returnee directors who are employees of the firm, divided by the total number of directors for each firm-year. Independent returnee directors denotes the number of returnee directors who are not employees of the firms, divided by the total number of directors. The sum of these two variables is the original Returnee directors. We then construct Independent local directors, which is the fraction of the number of nonreturnee directors who are not the firm’s employees divided by the total number of directors. In the regressions of tax avoidance, we replace the original Returnee directors variable with Independent returnee directors, Independent local directors, and Nonindependent returnee directors and reestimate the regressions. The results are reported in Panel D of Table 3. The coefficient of Independent returnee directors at the 10th percentile (0.1846) is positive, and that at the 90th percentile (−0.2385) is negative. The difference in the coefficients between the 10th and 90th percentiles is statistically significant, with a p value of .0000. The difference in the coefficients of Nonindependent returnee directors at the 90th quantile (−0.0303) and the 10th quantile (0.0184) is not statistically different, with a p value of .4949. Independent returnee directors positively affects tax avoidance at low levels of tax avoidance and a negative effect on tax avoidance at high levels of tax avoidance. Meanwhile, Nonindependent returnee directors exerts an insignificant effect at almost all levels of tax avoidance. F tests in the middle of this table show that these differences in effects are statistically significant for most quantiles.

The results also show that the effects of Independent returnee directors and Independent local directors have similar trends. Both these two groups of directors increase tax avoidance when it is relatively low and decrease tax avoidance when it is relatively high. The magnitudes of the effects of Independent returnee directors are larger than those of Independent local directors. In other words, the effects of Independent returnee directors are more positive (negative) than those of Independent local directors when the levels of tax avoidance are low (high). F tests in the middle of this table show that these differences in effects are statistically significant for most quantiles. If Independent local directors demonstrates independence in the conventional sense, these results suggest that Independent returnee directors matters beyond the independence of directors in the conventional sense. Overall, our results emphasize that returnee directors’ independence can potentially shift the level of tax avoidance. These results support H2d.

Robustness

We conduct robustness checks of our analysis by using several methods. First, we verify our findings’ robustness by replacing Residual book-tax gap with an alternative measure of tax avoidance, ETR, which is the ratio of income tax expenses divided by the adjusted total profits.¹² When the value of this variable is lower, the firm’s tax avoidance practice is more severe. Second, we exclude firms that have not appointed a returnee director so that firms remaining in the control group likely experience similar shocks as firms that have already hired directors with foreign experience. We also restrict the sample to firms that exist through the whole sample period and thus construct a well-balanced panel data to exclude the influences of the firms’ entries and exits and improve the estimators’ consistency. Third, to alleviate the potential influence of unobserved political connections on our findings (Adhikari et al., 2006; Tang, 2016; Wu et al., 2012), we add Political CEO as a control variable and reestimate the regression.¹³

Finally, because we view tax avoidance as one type of risky corporate activity, the relation between Returnee directors and tax avoidance can be similar to the relation between Returnee directors and other risk corporate activities. We consider this possibility by measuring the capital expenditure of the firm (CAPEX). Firms with more returnee directors are more likely to increase CAPEX when it is too low and decrease it when it is too high. The results suggest that directors with foreign experience act as effective monitors in other risky decisions of the firms, reinforcing our previous results on tax avoidance. The results of the above robustness tests are in IA Table 8 in the Internet Appendix. The results are robust.

Concluding Remarks

This study asks whether and how mobile individuals, such as returnee directors, may create or fortify linkages through which corporate tax avoidance practices of some countries influence other countries’ practices. We argue that returnee directors can lead firms to increase tax avoidance when the level of tax avoidance is too low and decrease tax avoidance when the level of tax avoidance is too high. To examine this argument empirically, we use quantile regressions to evaluate the relation between returnee directors and tax avoidance in different parts of tax avoidance distribution. We also use a natural experiment from China. Starting in the 1990s, China introduced provincial policies to attract returnees, which provided an exogenous shock to the supply of returnee talents. Using IV quantile regression, we show that firms with more returnee directors tend to increase tax avoidance when tax avoidance is well below its industry-size-year mean and decrease tax avoidance when tax avoidance is well above its industry-size-year mean.

Further exploration shows that the relation is stronger for SOEs than for non-SOEs and stronger for returnee directors who obtained MBA degrees, returnee directors with a background in accounting or auditing, or independent returnee directors than for other returnee directors. Overall, our results support the idea that returning corporate elites transmit corporate tax avoidance practices from abroad to companies in their countries of origin. We emphasize the importance of the board’s role in monitoring the firm’s tax avoidance practice. Because it is difficult to rule out all possible confounding factors and selection issues, it is possible that those factors or issues partly drive our results.

Supplemental Material

sj-pdf-1-jaf-10.1177_0148558X211017356 – Supplemental material for Returnee Directors and Corporate Tax Avoidance

Supplemental material, sj-pdf-1-jaf-10.1177_0148558X211017356 for Returnee Directors and Corporate Tax Avoidance by Jia Chen, Dongjie Chen, Li Liu and Zhong Wang in Journal of Accounting, Auditing & Finance

Footnotes

Acknowledgements

We are grateful to Liyan Han, Chunxin Jia, Jane Liu, Yizong Li, Yaping Wang, Yunhong Yang, Zhigang Zheng, the editors of Journal of Accounting, Auditing & Finance, two anonymous referees, and seminar at Guanghua School of Management of Peking University and participants of 2017 Journal of Accounting, Auditing & Finance Conference.

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: We thank the National Natural Science Foundation of China for financial support with grant number 71403010. All errors are our own.

ORCID iD

Jia Chen

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Abadie

Angrist

Imbens

(2002). Instrumental variables estimates of the effect of subsidized training on the quantiles of trainee earnings. Econometrica, 70(1), 91–117.

Adhikari

Derashid

Zhang

(2006). Public policy, political connections, and effective tax rates: Longitudinal evidence from Malaysia. Journal of Accounting and Public Policy, 25(5), 574–595.

Angrist

J. D.

Pischke

J. S.

(2008). Mostly harmless econometrics. Princeton University Press.

Armstrong

C. S.

Blouin

J. L.

Jagolinzer

A. D.

Larcker

D. F.

(2015). Corporate governance, incentives, and tax avoidance. Journal of Accounting and Economics, 60(1), 1–17.

Armstrong

C. S.

Blouin

J. L.

Larcker

D. F.

(2012). The incentives for tax planning. Journal of Accounting and Economics, 52(1), 391–411.

Bame-Aldred

C. W.

Cullen

J. B.

Martin

K. D.

Parboteeah

K. P.

(2013). National culture and firm-level tax evasion. Journal of Business Research, 66(3), 390–396.

Blouin

J. L.

Collins

J. H.

Shackelford

D. A.

(2005). Does acquisition by non-U.S. shareholders cause U.S. firms to pay less tax? Journal of the American Taxation Association, 27(1), 25–38.

Braguinsky

Mityakov

(2015). Foreign corporations and the culture of transparency: Evidence from Russian administrative data. Journal of Financial Economics, 117(1), 139–164.

Cai

Liu

(2009). Competition and corporate tax avoidance: Evidence from Chinese industrial firms. The Economic Journal, 119(537), 764–795.

10.

Chernozhukov

Hansen

(2005). An IV model of quantile treatment effects. Econometrica, 73(1), 245–261.

11.

Chesher

(2003). Identification in nonseparable models. Econometrica, 71(5), 1405–1441.

12.

Chyz

J. A.

Gaertner

F. B.

(2018). Can paying “too much” or “too little” tax contribute to forced CEO turnover? Accounting Review, 93(1), 103–130.

13.

Clarke

(2006). The independent director in Chinese corporate governance. The Delaware Journal of Corporate Law, 31(1), 125–228.

14.

DeBacker

Heim

B. T.

Tran

(2015). Importing corruption culture from overseas: Evidence from corporate tax evasion in the United States. Journal of Financial Economics, 117(1), 122–138.

15.

Desai

M. A.

(2003). The divergence between book income and tax income. Tax Policy and the Economy, 17(17), 169–206.

16.

Desai

M. A.

Dharmapala

(2006). Corporate tax avoidance and high-powered incentives. Journal of Financial Economics, 79(1), 145–179.

17.

Desai

M. A.

Dharmapala

(2009). Earnings management, corporate tax shelters, and book-tax alignment. National Tax Journal, 62(1), 169–186.

18.

Dyreng

S. D.

Hanlon

Maydew

E. L.

(2010). The effects of executives on corporate tax avoidance. The Accounting Review, 85(4), 1163–1189.

19.

Frank

M. M.

Lynch

L. J.

Rego

S. O.

(2009). Tax reporting aggressiveness and its relation to aggressive financial reporting. The Accounting Review, 84(2), 467–496.

20.

Frölich

Melly

(2013). Unconditional quantile treatment effects under endogeneity. Journal of Business & Economic Statistics, 31(3), 346–357.

21.

Giannetti

Liao

(2015). The brain gain of corporate boards: Evidence from China. The Journal of Finance, 70(4), 1629–1682.

22.

Gul

F. A.

Kim

J. B.

Qiu

A. A.

(2010). Ownership concentration, foreign shareholding, audit quality, and stock price synchronicity: Evidence from China. Journal of Financial Economics, 95(3), 425–442.

23.

Hanlon

(2005). The persistence and pricing of earnings, accruals, and cash flows when firms have large book-tax differences. Journal of the American Taxation Association, 80(1), 137–166.

24.

Hanlon

Heitzman

(2010). A review of tax research. Journal of Accounting and Economics, 50(2–3), 127–178.

25.

Healy

P. M.

(1985). The effect of bonus schemes on accounting decisions. Journal of Accounting and Economics, 7(1), 85–107.

26.

Jiang

Wan

Zhao

(2016). Reputation concerns of independent directors: Evidence from individual director voting. Review of Financial Studies, 29(3), 655–696.

27.

Kaplan

D. M.

Sun

(2017). Smoothed estimating equations for instrumental variables quantile regression. Econometric Theory, 33(1), 105–157.

28.

Knyazeva

Masulis

R. W.

(2013). The supply of corporate directors and board independence. Review of Financial Studies, 26(6), 1561–1605.

29.

Lanis

Richardson

(2011). The effect of board of director composition on corporate tax aggressiveness. Journal of Accounting and Public Policy, 30(1), 50–70.

30.

Lee

(2007). Endogeneity in quantile regression models: A control function approach. Journal of Econometrics, 141(2), 1131–1158.

31.

Lev

Nissim

(2004). Taxable income, future earnings, and equity values. Accounting Review, 79(4), 1039–1074.

32.

Manzon

Plesko

G. A.

(2002). The relation between financial and tax reporting measures of income. Tax Law Review, 55(2), 175–214.

33.

Mao

Zhou

Wang

(2012). The Governance Effect of Compensation Committee on Compensation Contract Effectiveness: Empirical Evidence from A-share Listed Companies 2002-2010, Review of Investment Studies, 31(9), 22–43.

34.

Minnick

Noga

(2010). Do corporate governance characteristics influence tax management? Journal of Corporate Finance, 16(5), 703–718.

35.

Naughton

(1995). Growing out of the plan. Cambridge University Press.

36.

Phillips

J. D.

(2003). Corporate tax-planning effectiveness: The role of compensation-based incentives. Accounting Review, 78(3), 847–874.

37.

Plesko

(2002). Reconciling corporation book and net tax income, tax years 1996-1998. SOI Bulletin, 21(4), 1–16.

38.

Powell

(2013). A new framework for estimation of quantile treatment effects: Nonseparable disturbance in the presence of covariates. http://www.rand.org/content/dam/rand/pubs/working_papers/WR800/WR824-1/RAND_WR824-1.pdf

39.

Rego

S. O.

Wilson

(2012). Equity risk incentives and corporate tax aggressiveness. Journal of Accounting Research, 50(3), 775–810.

40.

Richardson

(2006). Determinants of tax evasion: A cross-country investigation. Journal of International Accounting Auditing and Taxation, 15(2), 150–169.

41.

Richardson

(2008). The relationship between culture and tax evasion across countries: Additional evidence and extensions. Journal of International Accounting Auditing and Taxation, 17(2), 67–78.

42.

Richardson

Taylor

Lanis

(2013). The impact of board of director oversight characteristics on corporate tax aggressiveness: An empirical analysis. Journal of Accounting and Public Policy, 32(3), 68–88.

43.

Shleifer

Vishny

R. W.

(1994). Politicians and firms. Quarterly Journal of Economics, 109(4), 995–1025.

44.

Slemrod

(2004). The economics of corporate tax selfishness. National Tax Journal, 57(4), 877–899.

45.

Statistics China. (2013). 2013 China statistical yearbook. China Statistics Press.

46.

Stickney

C. P.

Mcgee

V. E.

(1982). Effective corporate tax rates the effect of size, capital intensity, leverage, and other factors. Journal of Accounting and Public Policy, 1(2), 125–152.

47.

Tang

T. Y. H.

(2016). Privatization, tunneling, and tax avoidance in Chinese SOEs. Asian Review of Accounting, 24(3), 274–294.

48.

Tsakumis

G. T.

Curatola

A. P.

Porcano

T. M.

(2007). The relation between national cultural dimensions and tax evasion. Journal of International Accounting Auditing and Taxation, 16(2), 131–147.

49.

Zhou

(2012). Political connections, tax benefits and firm performance: Evidence from China. Journal of Accounting and Public Policy, 31(3), 277–300.

50.

Zimmerman

J. L.

(1983). Taxes and firm size. Journal of Accounting and Economics, 5(2), 119–149.

51.

Zweig

(2006). Competing for talent: China’s strategies to reverse the brain drain. International Labour Review, 145(1–2), 65–90.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.70 MB