Earnings quality and travel convenience in China for non-local independent directors with accounting expertise

ABSTRACT

In the vast territory of China, travel convenience may be an important external influence on independent directors in their governance role. In this paper, the relationship between travel convenience and earnings quality is examined from the perspective of a non-local independent director with accounting expertise. We define travel convenience as existing where the high-speed railway takes less than one hour to travel from the location of the company’s office to the independent director’s home location. The study finds that travel convenience is positively associated with earnings quality of a company whose independent directors with accounting expertise are non-local. Earnings quality of companies whose independent directors with accounting expertise are local is higher than that of companies whose independent directors with accounting expertise are non-local and for whom travel is inconvenient. There is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors with accounting expertise are non-local but for whom travel is convenient. Further study shows that this conclusion is still valid if we define travel convenience as existing when the geographical distance between the company and the non-local independent directors’ home location is less than 200 km or 100 km. This indicates that travel convenience enhances the governance effect of non-local independent directors and improves the company’s earnings quality.

KEYWORDS:

• Earnings quality
• non-local
• independent directors with accounting expertise
• travel convenience

1. Introduction

In 2001, the China Securities Regulatory Commission issued the Guiding Opinion about Building Independent Directors in the Listed Company (hereinafter referred to as the Guiding Opinion), which signalled that the independent director system was formally enforced. The Guiding Opinion specifies that the independent directors shall give independent opinions on related party transactions of listed companies, disclosure of important matters and conduct harmful to the interests of minority shareholders. The purpose of introducing an independent director system into listed companies is exerting independent directors’ supervision function (Fama & Jensen, 1983; Wang, 2007; Zhao, Zeng, & Tan, 2008), and protecting the interests of shareholders, especially those of the minority shareholders. As an important system of arrangement of corporate governance, domestic and overseas scholars have undertaken active research on the governance effect of independent directors. The previous studies focus on the relationship between the independent directors and corporate performance (Agrawal & Knoeber, 1996, 2001; Bhagat & Black, 2002; Li & Lai, 2004; Rosenstein & Wyatt, 1990; Tang & Luo, 2005). Since then, scholars have analysed various factors that influence the effect of independent directors, such as the proportion of independent directors (Lu & Hu, 2015; Peasnell, Pope, & Young, 2005; Wang, 2007), background (Hu & Tang, 2008; Hu, Liu, & Ji, 2016; Xie, Iii, & Dadalt, 2003), reputation (Tan, Zheng, & Zhou, 2006; Huang, Liu, & Ma, 2016), director network (Chen, 2012; Wang & Hu, 2014), and has produced comparatively abundant research achievements. Chinese independent directors are often part-time and have their own full-time jobs. According to the living habits of Chinese people, the city where the independent directors work full-time is generally the city where the family is located. According to Table 2, nearly half of the independent directors’ homes are not located in the same city as the listed company for which they act as independent directors. In recent years, with the rise of New Geographic Economics, some scholars have begun to pay attention to the influence of geographical position (Dong, 2016; Liu, Li, & Sun, 2015; Sun & Liu, 2014; Zhou & Liu, 2016). Some scholars believe that the local independent directors who have information advantages play an important role in the inhibition of earnings management (Huang, Li, & Guo, 2016), while other scholars point out that, compared with local independent directors, non-local independent directors are less likely to conspire with the management, so they can supervise management teams better and reduce the level of earnings management (Jiang & Li, 2011). Some scholars believe that the local independent directors with accounting expertise have no significant effect on the earnings management of listed companies (Wang & Zhang, 2010). With the emergence of the high-speed railway, scholars begin to realise that convenient transportation can weaken the negative impact of geographical distance by enhancing the accuracy of information transmission (Atack, Jaremski, & Rousseau, 2014; Giroud, 2013). The multi-disciplinary research on geographical position, corporate governance and corporate finance has gained increasing attention from scholars, and has progressed. From existing literature, researchers divide independent directors into two categories from geographical position, such as local and non-local, and study the difference in corporate governance effects between them. As a result they can observe the influence of geographical position on the function of independent directors. However, with the rapid development of transportation, especially with many high-speed railways constructed and put into operation, convenient transportation has reduced the temporal distance dramatically, and the one-city effect has appeared. Such travel convenience can give non-local independent directors more advantages over time, energy and information, enhance the governance effect of non-local independent directors and protect the minority shareholders. Therefore, this paper investigates the governance effect from the perspectives of the travel convenience of non-local independent directors.

Table 1. Definitions of the variables.

Variable	Definitions of the variables
ABSACC	The absolute value of accrued earnings management, which is calculated using the modified Jones Model.
CONVENIENCE	A dummy variable that takes on the value of 1 if it takes no more than one hour from the home location of non-local independent directors with accounting expertise to the company by high-speed railway, and 0 otherwise.
LOCAL	A dummy variable that takes on the value of 1 if the home location of the independent directors with accounting expertise is in the same city as the company’s location, and 0 otherwise.
STATE	If the actual controller is the government, it is defined as a state-owned enterprise and takes the value of 1, and 0 otherwise.
LOSS	A dummy variable that takes the value of 1 if the profit of this year is negative and 0 otherwise.
INVREC	The flexible earnings management is equal to total inventory and accounts receivables divided by total assets.
SIZE	The company’s size, measured as the natural log of total assets.
ROA	Measured as net income divided by total assets.
LEV	Financial leverage, measured as debt divided by total assets.
GROWTH	The enterprise growth, which is represented by the revenue growth rate.
BIGAUDIT	A dummy variable that takes on the value of 1 if the listed company is audited by Big Four, and 0 otherwise.
OPINION	A dummy variable that takes on the value of 1 for a nonstandard auditing opinion, and 0 otherwise.
BOARDSIZE	The board size is equal to the number of directors.
INDEPENDENCE	The independence of the board is the proportion of independent directors.
EXSHR	A dummy variable that takes on the value of 1 if there is management shareholding, and 0 otherwise.
INSTITUTIONAL	The proportion of institutional investors.
TOP1	Share concentration is the percentage of shares held by the largest shareholder at the end of the year.
M/B	Market-to-book ratio, measured as the market value of the company divided by the book value.
SUSPECT	The motive for earnings management is a dummy variable that takes on the value of 1 if ROE is between 0 and 0.001, and 0 otherwise.
ATO	Total asset turnover rate.

Table 2. Descriptive statistics of the main variables.

Variable	Mean	STD	Median	Min	Max	N
ABSACC	0.1569	0.1960	0.1100	0.0016	2.2164	6249
CONVENIENCE	0.1994	0.3996	0	0	1	2818
LOCAL	0.5490	0.4976	1	0	1	6249
STATE	0.4503	0.4976	0	0	1	6249
LOSS	0.0856	0.2798	0	0	1	6249
INVREC	0.2746	0.1763	0.2488	0.0045	0.8104	6249
SIZE	22.1052	1.2546	21.9309	19.1512	26.2455	6249
ROA	0.0429	0.0529	0.0364	–0.1795	0.2233	6249
LEV	0.4483	0.2145	0.4491	0.0454	1.0648	6249
GROWTH	0.1735	0.4517	0.1030	–0.5822	4.4291	6249
BIGAUDIT	0.0554	0.2287	0	0	1	6249
OPINION	0.0277	0.1641	0	0	1	6249
BOARDSIZE	8.7606	1.7144	9	5	15	6249
INDEPENDENCE	0.3725	0.0541	0.3333	0.2857	0.6000	6249
EXSHR	0.7411	0.4381	1	0	1	6249
INSTITUTIONAL	0.0672	0.0901	0.0406	0.0014	0.5847	6249
TOP1	0.3626	0.1530	0.3448	0.0841	0.7695	6249
M/B	0.9886	0.9984	0.6536	0	7.7634	6249
SUSPECT	0.0018	0.0419	0	0	1	6249
ATO	0.6881	0.4973	0.5655	0.0537	3.1872	6249

The better the quality of earnings information, the more decision-making information it can provide to the investor (Dechow, Ge, & Schrand, 2010). The high-quality information the company provides to the investor plays a role in protecting the investor. Many scholars have studied the governance effect of having independent directors in the board and the audit committee, from the perspectives of earnings quality (Huang et al., 2016; Klein, 2002; Xie et al., 2003). Built on such prior research, this paper has chosen earnings quality as the variable to test the governance effect of independent directors.

The professionalism and independence of independent directors are important elements that influence their governance effects (Fama & Jensen, 1983; Huang et al., 2016). The US, in the Sarbanes-Oxley Act of 2002, required that there must be independent directors with accounting expertise in a listed company. In 2002, the China Securities Regulatory Commission promulgated the Listed Company Governing Rules, which specified that there must be independent directors with accounting expertise serving on the board of directors. To meet regulatory needs, all listed companies in China have engaged independent directors with accounting expertise. Compared with independent directors with other backgrounds, independent directors with accounting expertise have more ability to apply professional comprehension and form judgement on the financial information disclosed by listed companies. In addition, independent directors with accounting expertise usually serve on the audit committee, and consequently play a more significant role in management of earnings quality. Previous studies have mainly focused on how earnings quality is influenced by independent directors’ work experience, the proportion of independent directors, and their reputation (Hu et al., 2008; Huang et al., 2016), but there are no studies on the influence of travel convenience. The existing studies usually research the micro factors that influence the independent directors’ governance effect, while there is seldom a study that analyses the external macro environmental factors, which also contribute to the independent directors’ governance effect. China has a vast territory, and China’s talent resources are unevenly distributed. A large number of talent resources are concentrated in major cities. Hence, listed companies located in small cities have to hire talent in major cities to serve as independent directors in order to meet regulatory needs. Consequently, the listed companies in China have many non-local independent directors (Sun & Liu, 2014). Apart from that, the travel conditions are one of the most important external circumstances non-local independent directors face when they carry out their duties, so the developing transportation construction in China will doubtless have a massive impact on non-local independent directors. How exactly will travel convenience contribute to the governance effects of non-local independent directors? This remains a question to be explored.

This paper takes as its sample the A-share listed companies in China between 2011 and 2015 which only have one independent director with accounting expertise. We research the influence on earnings quality of travel convenience for a non-local independent director with accounting expertise. The study finds that travel convenience is positively associated with the earnings quality of a company having a non-local independent director with accounting expertise. Earnings quality of companies whose independent directors with accounting expertise are local is higher than that of companies whose independent directors with accounting expertise are non-local and whose travel is inconvenient. There is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors with accounting expertise are non-local but for whom travel is convenient.

There are three main theoretical contributions of this paper: first, we examine the influence of travel convenience on earnings management effects, as a novel factor that supports developing the field of study of independent directors’ management. The results of this study show that the travel convenience has an influence on the independent directors’ governance effect. Second, this study broadens the research perspective concerning the influence of geographical position on the function of independent directors. Prior literature mainly focuses on local and non-local independent directors, and their influences on governance. However as the travel conditions become more and more convenient, some places that may appear to be at a distance from each other become linked in a one-city effect. Based on the differentiation of non-local independent directors related to travel convenience, this paper deepens the research on the influence of travel convenience on the governance function of non-local independent directors. Third, this paper enriches research conclusions on the influence of the geographical position of the independent director on earnings quality. According to prior research, the local independent directors play a more important role in the inhibition of earnings management than the non-local ones. However, this research shows that travel convenience can help the non-local independent directors inhibit earnings management. Finally, we find there is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors with accounting expertise are non-local but for whom travel is convenient.

The remainder of this paper is as follows: Section 2 sets out the present situation of high-speed railway in China and the research hypotheses; Section 3 explains the sources of sample and data, the designs of models and the explanations of variables. Section 4 provides descriptive statistics, empirical results, and further tests. Section 5 describes robustness tests and Section 6 presents the conclusion.

2. Theoretical analysis and research hypothesis

2.1. The current situation of high-speed railway in china

The high-speed railway in China has developed rapidly since the Beijing–Tianjin inter-city railway line, which was the first high-speed railway with a speed up to 350 km per hour, was completed on 1 August 2008. China is a vast territory, and different regions have considerably different population densities; China has formed its own human geography environment with a large demand for people and logistics. Since the beginning of twenty-first century, with the support of national policy, China’s transportation has developed rapidly. The concrete developments are as follows: first, according to the mid- to long-term railway network plan, ‘11th five-year plan’ and ‘12th five-year plan’, the conformation of a dedicated passenger network mainly constituted by a four-vertical-and-four-horizontal fast-track¹ railway has been extensively progressed. A series of high-speed railways completed (such as the Beijing–Tianjin railway, Beijing–Shanghai railway, Beijing–Guangdong railway, Haerbin–Dalian railway, Shanghai–Nanjing railway) with a design speed of 350 km per hour and advanced world level features, has developed a sound technological management system of a high-speed railway. Second, the National Plan on New Urbanization (2014–2020) released in 2014 puts forward plans that improve the construction of city clusters in eastern China, such as the Beijing–Tianjin–Hebei region, the Yangtze River Delta and the Pearl River Delta. The plan also requires enterprises to foster and develop new city clusters in central and western China, such as the Chengdu–Chongqing region, the Central China Plain and the middle reaches of the Yangtze River, and finally forms an intercity-railway-based, motorway-based transportation network. Third, a ‘mid-long term railway network plan (2016-2030)’ revised in 2016 proposes to build a high-speed network, which is mainly based on eight-vertical-eight-horizontal,² and the inter-city railway as a supplement. This means the achievement of one to four hours’ travel radius³ within city clusters. The ‘13th Five-Year’ Plan points out that we should make great efforts to develop an inter-city railway, a regional railway in an urbanization region, and build one to two hours’ travel radius among centre cities and surrounding cities, and an hour’s travel radius between city centre and surrounding towns. According to statistics given by China Rail Transit Network, by the end of 2015, we have over 2000 EMUs, the vehicle density is about 0.93 vehicles/km, and the length of high-speed railways in service reaches 19,000 km, and over 8800 km of railway lines were put into operation, ranking first in the world. In conclusion, China’s transportation industry has developed rapidly, over the past few years. Meanwhile, the build of high-speed railways greatly promotes efficiency in logistics and flows among city clusters, and finally a one-city effect has appeared, meaning that the city clusters effectively form one city. Consequently, it is more meaningful to study the influence of travel convenience on the governance effect of non-local independent directors in the context of China.

2.2. Research hypotheses

Some researchers point out that the reputation (Fama & Jensen, 1983; Luo, 2014), specialisation (Cao & Chen, 2008; Xie & Yi, 2014), information advantage (Fu & Xia, 2014; Niu & Zhao, 2012), energy (Tan et al., 2003; Wang, Wu, & Xu, 2016) and independence (Liang & Zeng, 2016; Wang, Zhou, & Chen, 2008) of the independent directors have a marked influence on their governance effect. Existing literatures on geographical proximity show that distance is the main source of the information asymmetry (Ragozzino, 2009). Investors, analysts and underwriters have sufficient information about local stocks, so their forecasting precision is preferable to others (Bae, Stulz, & Tan, 2005; Baik, Kang, & Kim, 2010; Loughran & Schultz, 2005; Malloy, 2005). Meanwhile, the less the distance between the investors and companies, the more frequent and convenient is their communication (Chhaochharia, Kumar, & Niessen-Ruenzi, 2012; Coval & Moskowitz, 1999; Giannini, Irvine, & Shu, 2011), and it also greatly reduces the possibility of manipulating financial statements because of the lower monitoring cost (Ayers, Ramalingegowda, & Yeung, 2011). In the specific case of the non-local independent directors, they are facing similar difficulties with non-local investors, analysts and underwriters. The geographical distance impacts the companies’ earnings quality by influencing the supervising costs and information asymmetries of independent directors (Huang et al., 2016). Therefore, this paper will study the influences of the travel convenience of non-local independent directors with accounting expertise on the earnings quality from the perspectives of information asymmetry and monitoring cost.

On the one hand, compared with the local independent directors with accounting expertise, the non-local independent directors have a higher degree of information asymmetry. However, the travel convenience can reduce the degree of information asymmetry by influencing the information’s quantity and quality collected by the non-local independent directors with accounting expertise. A social network promotes the circulation of information and resource (Peng & Liao, 2008), and provides the actors with information, knowledge and strategic resources required by economic decision (Chen, 2012). The local independent directors have social network in their area, so apart from publicly available information (financial reports, analyst reports and media reports), they can get inside information about a company’s operation through the staff, customers, suppliers, relatives, friends and through other ways, which can help them to govern (Coval & Moskowitz, 1999). Next, the local independent directors can utilise their advantages of geography to have direct contact with executives and core staff for information. Not only are the geographical advantages beneficial for them to observe the company directly and consistently, to know the changes of environment in time, but they understand every part of the company as well. Since the above information is from face-to-face communication and direct observation (Petersen, 2004; Stein, 2002), independent directors are likely to understand the internal condition of the enterprise. Finally, the geographical distance has an effect on the efficiency of information transfer. Long-distance transmission of information can not only influence the timeliness of information, but also lead to distortion and incomplete provision (Dong, 2016), reducing the reliability of information. However, travel convenience can reduce the non-local independent directors’ information asymmetry. The traffic technology decides the intensity and scope of the spatial interaction, and it is also one of the most important factors that changes human economic activities. Convenient transportation promotes efficiency in logistics, flows and capital, and it is beneficial to enhance the cooperation and linkages between different regions. This kind of cooperation and linkages between different regions will definitely bring abundant information flows, providing much knowledge and information for decision-making. Travel convenience can shorten the time for long-distance transmission, and enhance the timelines and reliability of information. Accordingly, travel convenience decreases the degree of information asymmetry by influencing the information quality and quantity of non-local independent directors with accounting expertise.

On the other hand, compared with the local independent directors with accounting expertise, the non-local independent directors have higher monitoring costs, which weakens their governance effects, but travel convenience can cut down the monitoring cost. First, they live far from the company, so when they participate in board decision-making and management decision-making, they will spend much time and money on travel (Quan & Chen, 2016). Dong (2016) thinks that non-local independent directors spend time booking a flight and on the flight. Second, non-local independent directors expend more energy on monitoring than local independent directors (Quan & Chen, 2016). Owing to the geographical advantage, local independent directors can visit companies frequently and conveniently (Huang et al., 2016), and monitor management in a timely manner. During a board meeting, local independent directors have adequate energy for the various resolutions (Zhou & Liu, 2016). However, this could be an issue with non-local independent directors. For one thing, they may be absent from a board meeting and fail to manage effectively because of the inconvenience of travel, as a result of which, the management are much more likely to carry out earning management (Hu & Tang, 2008; Jiang & Li, 2011); for another, because of the limited time and energy (Tan et al., 2003), long travel time and new surroundings will cost significant amounts of this limited time and energy, so non-local independent directors spend less energy and time on the company’s operations, which weakens their management effectiveness (Tang & Luo, 2005; Wang, 2007). However, the travel convenience can make it more convenient and efficient for non-local independent directors with accounting expertise to participate in board decision-making and investigation, cutting down the cost of time and transportation. Apart from that, travel convenience makes their journey shorter and more comfortable, reduces the waste of their energy, and consequently, they have more time and energy to supervise the companies. Hence, travel convenience is beneficial for the non-local independent directors with accounting expertise to play the role in corporate governance, and then monitor the companies to enhance the earnings quality. We propose Hypothesis 1.

H1:

Travel convenience is positively associated with earnings quality of a company which has non-local independent directors with accounting expertise.

Prior research indicates that convenient transportation converts spatial distance into time distance (Wang, Zhu, & Wang, 2015), and the construction and operation of the high-speed railway has brought a time-space compression effect (Hall, 2009), which does not change the spatial distance but narrows the time distance, even causing a one-city effect (Liang, 2010). The existing studies show that the construction of the high-speed railway greatly improves the convenience of investors and underwriters to investigate the company on-the-spot (Huang et al., 2016) and communicate with each other (Atack et al., 2014; Hornung, 2015), reducing information asymmetry. The one-city effect, which results from travel convenience, enhances the intensity and scope of the interaction between different cities, and promotes the development of regional economic activities by integration (Wang, Jiao, & Jin, 2014). The one-city effect cuts down the non-local independent directors’ cost of time, energy and transportation. Meanwhile, it weakens the negative impact of geographical distance on information transmission. As a result, the one-city effect makes the local independent directors’ strengths of information asymmetry and monitoring cost less obvious compared with non-local ones, and so there may also be less difference in earnings quality. On the basis of this analysis, we propose Hypothesis 2.

H2:

There is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors are non-local but for whom travel is convenient.

3. Research design

3.1. Sample and data

This paper chooses as the research sample the A-share listed companies in China between 2011 and 2015 that only have one accounting independent director. We limit the sample interval to since 2011, mainly because from 30 June 2011, the Beijing–Shanghai high-speed railway was put into use and the high-speed railway began to have a wide coverage, which at the north extends to the Beijing–Tianjin–Hebei region and at the south to the Yangtze River Delta. Although China had a high-speed railway before, it only covered a few places, so the convenience for non-local independent directors was not enough and the conclusion is not so sound. In view of this, we select data from 2011 to 2015 as a sample. The criteria for selecting samples are as follows: (1) the financial data of the finance and insurance industry, which is quite different from others, should be eliminated; (2) the data of ST companies are removed;⁴ (3) incomplete or flawed data are removed; (4) to minimise the influence of outliers, we winsorise all continuous variables at the first and 99th percentiles. Finally, we obtain 6,249 observations.

Data on the registered address, home location and the independent directors’ related data of the listed companies come from manual collection by the authors. The data about the high-speed railway, which links the home location of non-local independent directors with accounting expertise to the home location of their companies is also collected manually, such as when it begins to operate and how long the journey takes,⁵ and other data from the CSMAR database. STATA 11.0 and Excel 2013 are adopted to analyse and process the data. Based on the general standard for the industrial classification provided by China’s Securities Regulatory Commission in 2012, the paper reclassifies the industry.⁶ Owing to limited space, the table of sample distribution for each year and industry is not listed.

3.2. Variables definition

3.2.1. Explained variables

In order to test the hypotheses, this paper chooses the earnings quality to be a measure of the independent directors’ governance effect. Based on the existing studies of earnings management (Huang et al., 2016; Li, Zhao, Xu, & Li, 2016; Sun, Liu, & Zhao, 2015), this paper uses the Modified Jones Mode (Dechow, Sloan, & Sweeney, 1995) to calculate the accrued earnings management to be the explained variables to measure the earnings quality of listed company.

According to the modified Jones Model, the regression method is used to estimate discretionary accruals (ACC) by sub-year and sub-industry samples, and the paper use ABSACC to measure the earnings management. The modified Jones Model is as follows:

(1) $\frac{T{A}_{i,t}}{A_{i,t-1}}={\alpha }_{0,i}+{\alpha }_{1,i}\left(\frac{1}{A_{i,t-1}}\right)+{\beta }_{1,i}\frac{\left(\mathrm{\Delta }{S}_{i,t}-\mathrm{\Delta }RE{C}_{i,t}\right)}{A_{i,t-1}}+{\beta }_{2,i}\left(\frac{PP{E}_{i,t}}{A_{i,t-1}}\right)+{\epsilon }_{i,t}$(1)

where TA_i,t is company i’s total accruals in year t, which equals the operating profit minus the net operating cash flow; ΔREC_i,t is the change in the ith company’s accounts receivable over year t; ΔS_i,t is the change in the ith company’s prime operating revenues during year t; PPE_i,t is the total original value of the ith company’s fixed assets at the end of year t; ${\epsilon }_{i,t}$is company i’s total assets at the end of year t–1; ε_i,t is the residual, the OLS method is used to estimate ${\epsilon }_{i,t}$ in model (1) by sub-year and sub-industry samples, and it also is ACC. The absolute value of it is ABSACC, which is the measurement of earnings management, the bigger the ABSACC, the higher the level of accruals earnings management and the worse the earnings quality of the company.

3.2.2. Explanatory variables

LOCAL shows whether the home location of an independent director with accounting expertise matches that of the company. LOCAL is a dummy variable that takes the value of 1 if the home location of the independent director with accounting expertise and the company is at the same prefecture-level city, and is 0 otherwise.

CONVENIENCE is the travel convenience of the non-local independent directors with accounting expertise, which shows how long it takes from the independent directors’ home location to the company by high-speed railway. CONVENIENCE is a dummy variable that takes the value of 1 if it takes no more than 1 hour, and is 0 otherwise.

3.2.3. Control variables

This paper draws on the studies by Wang and Wang (2007), Jiang, Zhu, and Tang (2013), Cheng, Zhong, and Yang (2015) and Huang et al. (2016) to use the nature of the company (STATE), the company’s size (SIZE), return-on-assets (ROA), financial leverage (LEV),the enterprise growth (GROWTH), the board size (BORADSIZE) and the board independence (INDEPENDENCE) as control variables.

3.3. Model design

With the method of multiple linear regression, this paper investigates the influence of travel convenience on the earnings quality of a company which has non-local independent directors with accounting expertise.

To test Hypothesis 1, the following empirical model is applied:

(2) $ABSACC={\alpha }_0+{\alpha }_{1}LOCAL+ControlVariables+INDUSTRY+YEAR+\epsilon$(2)

(3) $ABSACC={\beta }_0+{\beta }_{1}CONVENIENCE+ControlVariables+INDUSTRY+YEAR+\epsilon$(3)

Definitions of variables are shown in Table 1.

The regression analysis of model (2) for the whole sample is used to investigate the influence of independent directors’ geographical location on the earnings management of the listed company. If the coefficient of LOCAL is significantly negative, it means local independent directors play a more important role in the inhibition of earnings management. Under this premise, we test model (3) using the data for which the LOCAL value is 0 (i.e. all non-local independent directors with accounting expertise). If the coefficient of CONVENIENCE is significantly negative, there is an inverse association of with travel convenience for the company which has non-local independent directors with accounting expertise. In other words, as expressed in Hypothesis 1, travel convenience is positively associated with the earnings quality of the company which has non-local independent directors with accounting expertise.

In order to test Hypothesis 2, we use model (2) for two regressions. In each regression we are comparing local directors with non-local directors, but we separate out the two different types of non-local directors. In the first regression analysis (Sample 1) we compare companies whose independent directors with accounting expertise are local (we label these as LOCAL=1) with companies whose independent directors with accounting expertise are non-local and for whom travel is inconvenient (we label these as LOCAL=0). In this sample, the benchmark is companies whose independent directors with accounting expertise are non-local and for whom travel is inconvenient. If the coefficient of LOCAL is significantly negative, it means that, compared with non-local independent directors for whom travel is inconvenient, the local independent directors play a more important role in the inhibition of earnings management.

In the second regression analysis (Sample 2) we compare companies whose independent directors with accounting expertise are local (again we label these as LOCAL=1) with companies whose independent directors with accounting expertise are non-local but for whom travel is convenient (we label these as LOCAL=0). In this sample, the benchmark is companies whose independent directors with accounting expertise are non-local and for whom travel is convenient. If the coefficient of LOCAL is not significant, it means there is no significant difference in the influence on the earnings quality of local independent directors with accounting expertise and non-local accounting expertise independent directors for whom travel is convenient.

Therefore, in these two samples, we compare companies whose independent directors with accounting expertise are local with two different benchmarks, and we expect the results will be different. In Sample 1, the coefficient of LOCAL is negative and statistically significant while in Sample 2 the coefficient of LOCAL is not significant.

4. Empirical results

4.1. Descriptive statistics

Table 2 provides descriptive statistics calculated by STATA 11.0 for our observations. As seen from Table 2, the mean value of LOCAL is 0.549, indicating that 54.9% of the sample engage local independent directors with accounting expertise, and 45.1% of the sample engage non-local independent directors with accounting expertise. The mean value of CONVENIENCE is 0.199, showing that 19.9% of the companies that engage non-local independent directors choose non-local independent directors for whom travel is convenient. The non-local independent directors for whom travel is not convenient account for almost 80% of the total. The descriptive statistics of other variables are shown in Table 2.

4.2. Regression test of Hypothesis 1

Hypothesis 1 predicts that travel convenience is positively associated with the earnings quality of the company, which has non-local independent directors with accounting expertise. Table 3 reports the regression results for the test of Hypothesis 1.

Table 3. Regression of travel convenience and the governance effect of non-local independent directors with accounting expertise.

	Explained variables; ABSACC
Variables	(1) Total sample	(2) Non-local sample
LOCAL	–0.011**
	(–2.457)
CONVENIENCE		-0.026***
		(-3.237)
STATE	–0.013**	-0.004
	(–2.44)	(-0.490)
LOSS	0.055***	0.052***
	(4.665)	(2.883)
INVREC	–0.166***	-0.171***
	(–7.483)	(-4.394)
SIZE	–0.048***	-0.059***
	(–11.452)	(-8.122)
ROA	–0.002	0.086
	(–0.021)	(0.532)
LEV	0.100***	0.140***
	(3.920)	(3.113)
GROWTH	0.143***	0.151***
	(8.855)	(6.614)
BIGAUDOT	0.021**	0.046***
	(2.552)	(3.621)
OPINION	0.144***	0.147***
	(4.965)	(3.804)
BOARDSIZE	0.001	0.001
	(0.766)	(0.288)
INDEPENDENCE	0.117**	0.199***
	(2.518)	(2.646)
EXSHR	–0.010*	-0.021**
	(–1.755)	(-2.325)
INSTITUTIONAL	–0.005	0.017
	(–0.223)	(0.431)
TOP1	0.075***	0.107***
	(4.516)	(4.213)
M/B	0.017***	0.014**
	(4.799)	(2.387)
SUSPECT	0.021	0.053
	(0.322)	(0.928)
ATO	0.068***	0.050***
	(7.084)	(3.846)
Industry & Year	Yes	Yes
Observations	6249	2818
F Value	11.237***	6.377***
Adjusted R^2	0.234	0.276

Note: ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively.

Column (1) shows the regression results of model (2) in the whole sample. The regression results show that after controlling for other variables, the coefficient on LOCAL remains negative and significant at the 5% level. Therefore, local independent directors have a better effect on inhibiting earnings management than non-local independent directors with accounting expertise, and consequently, their companies have better earnings quality. Then we employ regression analyses in the non-local independent directors sample and the results are listed in Column (2). The regression results show that after controlling for other variables, the coefficient on CONVENIENCE remains negative and significant at the 1% level, indicating that compared with non-local independent directors with inconvenient travel, the non-local independent directors for whom travel is convenient play a more important role in the inhibition of earnings management, and the travel convenience enhances the governance effect of non-local independent directors with accounting expertise, which improves the companies’ earnings quality. The results in Table 3 support Hypothesis 1.

4.3. Regression test of Hypothesis 2

Hypothesis 2 predicts that there is no significant difference in the company’s earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors are not local, but for whom travel is convenient. We describe this as a one-city effect. Table 4 reports the regression results for the test of Hypothesis 2. For this test we construct Sample 1 by combining companies whose independent directors with accounting expertise are local (labelled as LOCAL=1) with those whose independent directors with accounting expertise are non-local but for whom travel is inconvenient (labelled as LOCAL=0). We contrast this with Sample 2 which contains companies whose independent directors with accounting expertise are local (LOCAL=1) and those whose independent directors with accounting expertise are non-local but for whom travel is convenient (LOCAL=0).

Table 4. Regression of one-city effect.

	Explained variables: ABSACC
Variables	(1) Sample1	(2) Sample2
LOCAL	–0.016***	0.009
	(–3.194)	(1.341)
STATE	–0.014**	–0.012**
	(–2.530)	(–2.107)
LOSS	0.051***	0.056***
	(4.052)	(4.223)
INVREC	0.167***	–0.157***
	(–7.061)	(–7.381)
SIZE	0.049***	–0.041***
	(–10.916)	(–8.789)
ROA	0.001	–0.057
	(–0.010)	(–0.572)
LEV	0.112***	0.050**
	(3.885)	(1.985)
GROWTH	0.137***	0.116***
	(8.234)	(7.692)
BIGAUDOT	0.020**	0.005
	(2.237)	(0.479)
OPINION	0.154***	0.106***
	(4.968)	(3.238)
BOARDSIZE	0.001	0.001
	(0.520)	(0.723)
INDEPENDENCE	0.118**	0.059
	(2.465)	(1.173)
EXSHR	–0.008	–0.000
	(–1.397)	(–0.019)
INSTITUTIONAL	–0.010	0.006
	(0.418)	(0.243)
TOP1	0.082***	0.041**
	(4.592)	(2.293)
M/B	0.017***	0.022***
	(4.307)	(4.948)
Suspect	0.018	0.026
	(0.256)	(0.291)
ATO	0.068***	0.073***
	(6.585)	(6.678)
Industry & Year	Yes	Yes
Observations	5687	3993
F Value	10.708***	8.697***
Adjusted R^2	0.226	0.251

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively.

Column (1) shows the regression results of model (2) in Sample 1. The regression results show that the coefficient on LOCAL remains negative and significant at the 1% level. Then we employ regression analysis in Sample 2. The results are reported in column (2), and the coefficient on LOCAL is not significant. To sum up, there is a significant difference between non-local accounting expertise independent directors for whom travel is inconvenient and local accounting expertise independent directors in the role of governance; however, the travel convenience of independent directors weakens this difference. That is to say, there is a one-city effect of non-local accounting expertise independent directors for whom travel is convenient. This one-city effect, which results from travel convenience, enhances the intensity and scope of the interaction between different cities. The one-city effect cuts down the non-local independent directors’ costs of time, energy and travel. Meanwhile, it weakens the negative impact of geographical distance on information transmission. As a result, the one-city effect makes the local independent directors’ strengths of information asymmetry and monitoring cost less obvious compared with non-local ones, so there may be less differences in governance effect. The results in Table 4 support Hypothesis 2.

4.4. Further tests

4.4.1. The convenience of highway travel

Given that the independent directors with accounting expertise who have short geographical distances may choose a car or bus ride rather than high-speed railway for a variety of reasons, we use geographical distances⁷ to measure the travel convenience, and Table 5 shows the regression results of model (2) and model (3). The speed limit on motorways in China is 120 km an hour, and the speed limit of national highway is 60–80 km an hour. Assuming that a 2 hour drive is generally accepted, we separately choose 100 km and 200 km as the dividing standards of travel convenience. CONVENIENCE takes on the value of 1 if the geographical distances between the home locations of independent directors with accounting expertise and the companies are no more than 200 km (100 km), and 0 otherwise.

Table 5. Regression of the convenience of highway travel.

	Panel A: Hypothesis 1		Panel B: Hypothesis 2
	(1) 200km (Non-local)	(2) 100km (Non-local)	(3) 200km (Sample2)	(4) 100km (Sample2)
Variables	ABSACC	ABSACC	ABSACC	ABSACC
CONVENIENCE	–0.027***	–0.019**
	(–3.844)	(–2.473)
LOCAL			0.006	0.006
			(1.212)	(0.955)
Control variables	Yes	Yes	Yes	Yes
Industry & Year	Yes	Yes	Yes	Yes
Observations	2818	2818	4474	3928
F Value	6.447***	6.301***	9.702***	8.895***
Adjusted R^2	0.277	0.275	0.242	0.218

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) T statistics appear in square brackets. (3) The results of control variables have been removed to save space.

According to the results in Table 5 Panel A, the coefficient on CONVENIENCE in Column (1) and Column (2) are both significantly negative, which shows that there is a negative correlation between the convenience of highway travel and the earnings management of a company which has non-local independent directors with accounting expertise. According to the results in Panel B, the coefficients on CONVENIENCE in Column (3) and Column (4) are non-significantly positive, which shows that when we take the highway into account, the difference in the governance effect between non-local independent directors of the travel convenience and local independent directors with accounting expertise is not obvious, and a one-city effect appears.

4.4.2. The convenience of air travel

Given that the non-local independent directors with accounting expertise who have long geographical distances may choose to fly rather than use other transportation modes, we manually collected the data about the air route and the shortest flying time between the home locations of independent directors with accounting expertise and the respective companies through the website (www.ctrip.com). Taking the fixed time cost⁸ of air travel into account, we use the direct flight time to measure the travel convenience (CONVENIENCE), and we use 2 hours to be the criterion of CONVENIENCE. CONVENIENCE takes on the value of 1 if the direct flight time between the home location of an independent director with accounting expertise and the company is less than or equal to 2 hours, and 0 otherwise.

Table 6 column (1) uses model (3). The sample includes companies whose independent directors with accounting expertise are non-local, companies whose independent directors with accounting expertise are non-local and travel convenient (CONVENIENCE=1). Model (2) for column (2), sample includes companies whose independent directors with accounting expertise are local (LOCAL=1) and companies whose independent directors with accounting expertise are non-local and travel is inconvenient (LOCAL=0). Column (3) also uses model (2), this sample includes companies whose independent directors with accounting expertise are local (LOCAL=1) and companies whose independent directors with accounting expertise are non-local and for whom travel is convenient (LOCAL=0). According to the results in Column (1), the coefficient on CONVENIENCE is 0.005, and is non-significantly positive, which shows that when we take air travel into account, whether or not the non-local independent directors with accounting expertise are of the convenient-travel type does not influence their abilities to restrain the listed companies’ earnings management. According to the results in Column (2) and Column (3), the coefficient on CONVENIENCE is negative and significant at the 5% level, which shows that the convenience of air travel does not cause a one-city effect.

Table 6. Regression of the convenience of air travel.

	Panel A: Hypothesis 1	Panel B: Hypothesis 2
	(1) Non-local	(2) Sample1	(3) Sample2
Variables	ABSACC	ABSACC	ABSACC
CONVENIENCE	0.005
	(0.565)
LOCAL		–0.010**	–0.018**
		(–1.999)	(–2.333)
Control variables	Yes	Yes	Yes
Industry & Year	Yes	Yes	Yes
Observations	2818	5606	4704
F Value	6.317***	10.523***	8.086***
Adjusted R^2	0.274	0.245	0.216

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) T statistics appear in square brackets. (3) The results of control variables have been removed to save space.

5. Robustness tests

5.1. Endogeneity tests

In order to solve the endogeneity problem in selecting independent directors, this paper use a two-stage regression of instrumental variables to redesign model (3). In the first-stage of regression, CONVENIENCE is regarded as the dependent variable, and the control variables in model (3) are regarded as exogenous variables. In particular, we use SAMECSQ as an exogenous variable, which is a dummy variable and shows whether the company and independent directors’ home location are located in the same urban agglomeration.⁹ The reason is that in order to reduce costs and facilitate communication, the listed companies prefer to choose the independent directors from the nearby areas. As a result, whether the company and independent directors’ home location are located in the same urban agglomeration (SAMECSQ) will obviously influence the company to choose the non-local independent directors of the travel convenience, but SAMECSQ is not in direct contact with the company’s earnings quality. It is reasonable to use SAMECSQ as an exogenous variable, which is a dummy variable and shows whether the company and independent directors’ home location are located in the same urban agglomeration. We get an inverse mills ratio (Lambda) through the first-stage of regression, and then in the second-stage, we added Lambda as an extra control variable and employ regression analysis on model (3).

According to the first-stage regression results in Column (1) of Table 7, the coefficient on SAMECSQ and CONVENIENCE is significantly positive at the 1% level, indicating that instrumental variables are reasonable. Column (2) shows the regression results of the second-stage, and it is easy to see that the coefficients on CONVENIENCE and ABSACC are significantly negative at the 1% level, which is consistent with Hypothesis 1. These results support the conclusion that after controlling for endogeneity problems, non-local independent directors with accounting expertise for whom travel is convenient have a better effect on inhibiting earnings management compared with those non-local independent directors with accounting expertise for whom travel is not convenient. The travel convenience improves the earnings quality of companies whose independent directors with accounting expertise are non-local.

Table 7. Two-stage regression of instrumental variables.

	(1) First-stage	(2) Second-stage
Variables	Explained variable: CONVENIECE	Explained variable: ABSACC
CONVENIENCE		–0.055***
		(–2.932)
SAMECSQ	1.482***
	(21.672)
STATE	0.132*	–0.004
	(1.738)	(–0.508)
LOSS	–0.191	0.051***
	(–1.389)	(3.465)
INVREC	–0.196	–0.172***
	(–0.791)	(–6.381)
SIZE	–0.072	–0.060***
	(–1.612)	(–12.483)
ROA	–0.526	0.079
	(–0.647)	(0.883)
LEV	–0.183	0.138***
	(–0.818)	(5.573)
GROWTH	0.011	0.151***
	(0.164)	(21.513)
BIGAUDOT	–0.481***	0.044***
	(–2.608)	(2.664)
OPINION	0.176	0.147***
	(0.934)	(7.111)
BOARDSIZE	0.011	0.001
	(0.465)	(0.256)
INDEPENDENCE	–0.207	0.197***
	(–0.308)	(2.718)
EXSHR	0.149*	–0.019**
	(1.899)	(–2.257)
INSTITUTIONAL	0.412	0.021
	(1.144)	(0.505)
TOP1
	–0.017	0.110***
	(–0.073)	(4.338)
M/B	0.018	0.014**
	(0.318)	(2.477)
Suspect	–0.433	0.053
	(–0.569)	(0.698)
ATO	–0.052	0.050***
	(–0.723)	(6.202)
Lambda		0.021*
		(1.764)
Industry & Year	Yes	Yes
Observations	2818	2818

Note: ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively.

We build a difference-in-difference (DID) model to further solve the endogeneity problem in selecting independent directors. We define CONVENIENCE as travel convenience for the independent directors with accounting expertise, which shows how long it takes from the independent directors’ home location to the company by high-speed railway. It is a dummy variable that takes the value of 1 if it takes no more than an hour, and 0 otherwise. However, many companies’ independent directors with accounting expertise change every year, and it causes CONVENIENCE to change as well, as a result of which the number of treatment groups is fewer (only 10). Compared with the former controlled group (about 250), it is difficult to build a difference-in-difference (DID) model.

5.2. Redefining variables

We redefine CONVENIENCE in order to make the result more robust. CONVENIENCE takes on the value of 1 if the shortest time of a high-speed railway between the home locations of an independent director with accounting expertise and the company is not more than 2 hours, and 0 otherwise. Table 8 shows the regression results of the retest of Hypotheses 1 and 2. According to the results in Panel A, the coefficient on CONVENIENCE in Column (1) is significantly negative, which supports Hypothesis 1; according to the results in Panel B, the coefficient on LOCAL in Column (2) is significantly negative, and the coefficient on LOCAL in Column (3) is not significantly positive, which is consistent with Hypothesis 2.

Table 8. Regression of redefining variables.

	Panel A：Hypothesis 1	Panel B：Hypothesis 2
Variables	(1) Non-local	(2) Sample1	(3) Sample2
	ABSACC	ABSACC	ABSACC
CONVENIENCE	–0.017**
	(–2.271)
LOCAL		–0.016***	0.003
		(–3.078)	(0.497)
Control variables	Yes	Yes	Yes
Industry & Year	Yes	Yes	Yes
Observations	2818	5533	4147
F Value	6.352***	10.379***	9.067***
Adjusted R^2	0.275	0.227	0.250

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) T statistics appear in square brackets. (3) The results of control variables have been removed to save space.

5.3 Period of service of independent directors with accounting expertise

Differences in the period of service by independent directors have negative and positive impacts on their governance effects (Chen & Xiang, 2017). On the one hand, with the extension of the period, independent directors with accounting expertise will master more financial and business information of the companies, as a result of which they will have better professional competence, and this will exert a positive impact on their governance effects. On the other hand, with the extension of the period, independent directors with accounting expertise seem to be closer to their companies’ executives, and they might lose their independence, increasing the possibility that independent directors conspire with their companies’ executives. Our research shows the convenience of a high-speed railway is positively associated with the earnings quality of the company which has non-local independent directors with accounting expertise, and a one-city effect appeared. So one of the most important questions is whether the length of service of independent directors with accounting expertise will affect the empirical results.

To answer this question, empirical models (4) and (5) are defined:

(4) $ABSACC={\lambda }_0+{\lambda }_{1}CONVENIENCE+{\lambda }_{2}SERVICE+{\lambda }_{3}CONVENIENCE\times SERVICE+ControlVariables+INDUSTRY+YEAR+$(4)

(5) $ABSACC=+{\delta }_{1}LOCA{L}_{\hspace{0.17em}}+{\delta }_{2}SERVICE+{\delta }_{3}LOCA{L}_{\hspace{0.17em}}\times SERVICE+ControlVariables+INDUSTRY+YEAR+\epsilon$(5)

Specific definitions of variables are shown in Table 1. SERVICE is a dummy variable that takes on the value of 1 if the period of service of independent directors with accounting expertise is more than 3 years and 0 otherwise.

Table 9 shows the regression results of model (4) for the non-local independent directors sample (LOCAL=0), the regression results of model (5) for the sample comprising non-local independent directors who experience travel inconvenience and local independent directors with accounting expertise (LOCAL=1), and the regression results of model (6) for the sample comprising non-local independent directors experiencing the travel convenience and local independent directors with accounting expertise (LOCAL=1). According to the results in Panel A, the coefficient on CONVENIENCE×SERVICE in Column (1) is non-significantly negative, which shows that the re-election of independent directors with accounting expertise has no impact on the relationship between their travel convenience and their governance effects. According to the results in Panel B, the coefficient on LOCAL×SERVICE in column (2) and column (3) are both non-significant, so that there is no evidence that the re-elections of independent directors with accounting expertise has an impact the on one-city effect. The results show that the term of independent directors with accounting expertise will not affect the empirical results.

Table 9. Regression of test given the term of independent directors with accounting expertise.

	Panel A: Hypothesis 1	Panel B: Hypothesis 2
	(1) Non-local	(2)Sample1	(3) Sample2
Variables	ABSACC	ABSACC	ABSACC
CONVENIENCE×SERVICE	–0.012
	(–0.833)
LOCAL×SERVICE		–0.006	0.003
		(–0.572)	(0.224)
Control variables	Yes	Yes	Yes
Industry & Year	Yes	Yes	Yes
Observations	2818	5687	3993
F Value	6.135***	10.314***	8.426***
Adjusted R^2	0.276	0.226	0.252

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) T statistics appear in square brackets. (3) The results of control variables have been removed to save space.

5.4. Different division standards for local and non-local independent directors

LOCAL takes on the value of 1 if the independent directors’ home location and the registered address of company are in same prefecture-level city and the value 0 otherwise, and then we retest Hypothesis 1 and Hypothesis 2, the results for the test are shown in Table 10. As is shown in Column (1), the coefficient on CONVENIENCE is significant and negative, supporting Hypothesis 1. Column (2) shows that the coefficient on LOCAL is positive but not significant, which is consistent with Hypothesis 2.

Table 10. Regression of test given the different division standards for local and non-local independent directors.

Variables	Explained Variables: ABSACC
	(1) Hypotheses 1 (Non-local)	(2) Hypotheses 2 (Sample2)
CONVENIENCE	–0.027***
	(–2.689)
LOCAL		0.006
		(0.835)
Control variables	Yes	Yes
Industry & Year	Yes	Yes
Observations	3015	3860
F Value	6.433***	7.986***
Adjusted R^2	0.290	0.200

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) The results of control variables have been removed to save space.

5.5 Selection criteria for the sample of independent directors

We further examine the influence of different selection criteria for the sample of independent directors with accounting expertise on our conclusion. Table 11 shows the regression results of model (2) and model (3) in the sample of independent directors with accounting expertise who have served in the company for more than 6 months and who served in the company on 31 December (the balance sheet date), respectively. The conclusion is not changed.

Table 11. Regression of the test given the different selection criteria for the sample of independent directors.

	Panel A: Hypothesis 1		Panel B: Hypothesis 2
	(1) more than 6 months	(2) on 31 December	(3) more than 6 months	(4) on 31 December
Variables	ABSACC	ABSACC	ABSACC	ABSACC
CONVENIENCE	–0.025***	–0.026***
	（–3.126）	（–3.253）
LOCAL			0.009	0.007
			（1.371）	（1.032）
Control variables	Yes	Yes	Yes	Yes
Industry & Year	Yes	Yes	Yes	Yes
Observations	2764	2781	3927	3946
F Value	6.413***	6.326***	8.612***	7.089***
Adjusted R^2	0.264	0.277	0.252	0.130

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) T statistics appear in square brackets. (3) The results of control variables have been removed to save space.

5.6 Test of the difference between the listed company’s home location and the registered address

We conduct a test on Hypotheses 1 and 2 by using the sample where the home location is the same as the registered address. Table 12 shows the regression results.

Table 12. Test of the difference between the home location and the registered address.

Variables	Explained variables: ABSACC
	(1) Hypotheses 1 (Non-local)	(2) Hypotheses 2 (Sample2)
COVENIENCE	–0.023***
	(–2.720)
LOCAL		0.006
		(0.905)
Control variables	Yes	Yes
Industry & Year	Yes	Yes
Observations	2599	3742
F Value	5.928***	7.425***
Adjusted R^2	0.247	0.211

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) The results of control variables have been removed to save space.

As is shown in Column (1), the coefficient on CONVENIENCE is significantly negative, which is consistent with Hypothesis 1. Column (2) shows that the coefficient on LOCAL is positive but not significant, supporting Hypothesis 2.

5.7. Testing for differences in sample size

Given that the sample size of non-local independent directors with convenient and inconvenient travel are quite different, we use propensity score matching (PSM) to re-estimate model (3), which is used in Hypothesis 1. The non-local independent directors with convenient travel are categorised as the experimental group (TREAT=1) while those with inconvenient travel are the reference group. Then we use gross domestic product (GDP) as the explanatory variable to build the first stage of PSM, which is used to match the sample. The matched sample is used in multiple regression analysis. Our reasons for using GDP are as follows: the city with the higher GDP has a higher level of economic and social development, and its travel arrangements are also more developed, which makes it more convenient to contact the outside. The GDP of the listed company’s location influences whether the company can hire the non-local independent directors for whom travel is convenient. Figure 1 shows that the distributions of kernel density estimation of matched samples are very close to each other, which means that the differences of other characteristics between these samples are very small.

Figure 1. Result of propensity score matching.

The regression results in Table 13 show that TREAT is significantly negatively associated with ABSACC at the 1% level, indicating that, compared with the non-local independent directors with accounting expertise who have inconvenient travel, those with convenient travel play a more important role in the inhibition of earnings management, and the travel convenience improves the earnings quality of the non-local independent directors’ company, which is consistent with Hypothesis 1.

Table 13. Regression after propensity score matching.

Variables	Explained variables: ABSACC
TREAT	–0.0409**
	(–2.44)
Control variables	Yes
Industry & Year	Yes
Observations	1124

Note: (1) ***, ** and * represent significance at the 1%, 5% and 10% levels, respectively. (2) The results of control variables have been removed to save space.

6. Conclusion

This study examines the association of a company’s earnings quality with the travel convenience of independent directors having accounting expertise. The results show that travel convenience is positively associated with the earnings quality of a company which has non-local independent directors with accounting expertise. Furthermore, the earnings quality of companies whose independent directors with accounting expertise are local is higher than that of companies whose independent directors with accounting expertise are non-local and for whom travel is inconvenient. There is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors with accounting expertise are non-local but for whom travel is convenient. The implications of this research are as follows: first, travel convenience can cut down the monitoring cost for non-local independent directors with accounting expertise and reduce information asymmetry. As a result, this can enhance the governance effect of the non-local independent directors with accounting expertise, and strengthen their supervision and governance over companies. Second, this paper finds that travel convenience is an important extrinsic factors in influencing how independent directors function in the governance of companies, providing a significant control variable for further study, and reducing the endogeneity problems. Third, earnings quality of a company whose independent directors with accounting expertise are local or who are non-local but experience convenient travel is better than that of a company which has independent directors with accounting expertise who are non-local and experience inconvenient travel. There is no significant difference in earnings quality between companies whose independent directors with accounting expertise are local and those whose independent directors are non-local but for whom travel is convenient. As a result, travel convenience is a significant factor to be considered when the company hires non-local independent directors with accounting expertise.

This paper has the following limitations and requires further study. First, according to the requirements of The Guidelines for Management of Listed Company, the listed company has to employ at least one independent director with accounting expertise, but there is no mandatory requirement regarding other expertise. Limited to the comparability of the companies, this paper only pays attention to the influence of travel convenience for the independent director with accounting expertise, and does not take into consideration independent directors with other backgrounds, such as law, or a technical background. Second, limited to the theme of this research, this paper only tests the effect of travel convenience effect on earnings quality, without testing the effects on other dimensions. Future studies can further test the influence of travel convenience on other dimensions of independent directors with other backgrounds.

Acknowledgements and funding

We appreciate constructive comments and valuable suggestions from anonymous reviewers, Professor Pauline Weetman (the English language editor), and acknowledge financial support from the National Social Science Foundation Projects (grant no. 18BGL070).

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ This description refers to a long-term railway network plan promulgated by the Ministry of Railways of the People’s Republic of China, featuring four lines running north–south and four running east–west, completed in 2015.

² The eight vertical and eight horizontal high-speed railway network is the description of the 2016 revision of planning of China’s high-speed railway network, meaning eight lines running north–south and eight running east–west.

³ The travel radius refers to the reachable range within a certain time, centred on a certain place.

⁴ According to China’s stock listing rules, special treatment will be given to the stock transactions of listed companies with abnormal financial or other conditions.

⁵ The data about when the high-speed railway begin to operate and how long the journey takes is obtained and analysed from websites including kyfw.12306.cn, www.gaotie and baike.baidu.com.

⁶ According to the code of the industry, we combine P, Q, R into the same category, and H and S into another. Given there are many sub-sectors of the manufacturing industry, we reclassify it into 13 sub-sectors. In addition, other industries are divided by industry code.

⁷ Geographical distance is calculated by longitude and latitude. The data are collected manually from Google Maps with the precision of prefecture-level city, and are calculated using Google distance calculator

⁸ Such as the service time of a security check at the airport and the time taken to travel from home to the airport.

⁹ There are three urban agglomerations referred to in this paper, namely the Beijing–Tianjin–Hebei urban agglomeration, the Pearl River Delta urban agglomeration and the Yangtze River Delta urban agglomeration. The specific city name is derived from the Baidu Encyclopedia.