Foreign Experienced CEOs’ and Financial Statement Comparability

ABSTRACT

This research examines whether the characteristics of top management teams are related to financial statement comparability. Considering the foreign experienced CEOs (FCEOs) in Chinese listed firms from 2005 to 2018, we show that it can increase financial statement comparability. We argue that firms led by FCEOs will enhance information environment and governance mechanisms by minimizing agency issues, consequently generating more comparable financial statements. Through further investigation, we find that the relationship between FCEOs and comparability is stronger when CEOs possess a financial and accounting background and when they have overseas professional experiences. fthe relationship is weaker with the existence of higher economic policy uncertainty. Finally, we identify financial reporting as an important channel that explains the relationship between FCEOs and comparability. Findings remain consistent after numerous robustness checks and supplementary investigations comprising lag of independent variables, generalized method of moment, instrumental variable approach, propensity score matching, and alternative comparability measures. Overall, the findings suggest that CEOs’ foreign experiences are associated with corporate outcomes.

KEYWORDS:

• CEOs’ foreign experiences
• financial statement comparability
• corporate governance
• China

JEL:

• G34
• H26

1. Introduction

Top management team (TMT) characteristics, especially of CEOs, substantially influence corporate decisions (Lai and Liu 2018; Zhang 2018). Despite scholarly recognition of CEOs’ traits, such as expertise, tenure, personality traits, duality, education, and gender, CEO foreign experiences are a few of the factors currently gaining research interest. This stream of literature claims that foreign experienced CEOs¹ (FCEOs) have a possible influenfgce on corporate decisions, such as compensation (Conyon et al. 2019), undertaking foreign IPOs (Duan, Hou, and Rees 2020), performance (Duan and Hou 2017), and corporate investment (Dai, Kong, and Liu 2018). However, one of the important issues that has not received scholarly attention is financial statement comparability (comparability, hereafter). Accordingly, the current study contributes to the literature by investigating how FCEOs affect comparability.

Comparability is a crucial element in financial reporting (International Accounting Standard Board 2010). Users of financial statements benefit from it given its role to facilitate understanding of accounting information and thus in coming up with informed decisions. Studies (e.g., De Franco, Kothari, and Verdi 2011; Zhang 2018) suggests that comparability enhances information quality while reducing acquisition and processing costs. Comparability can correspondingly diminish information asymmetry, which leads to higher innovation efficiency (Chircop et al. 2020); more efficient resource allocation (Kim et al. 2020); higher stock price information quality (Choi et al. 2019); greater accounting information’s relevance (Kim, Kim, and Musa 2018); lower cost of capital (Imhof, Seavey, and Smith 2017); and higher analyst coverage and forecast accuracy and lower forecast dispersion (De Franco, Kothari, and Verdi 2011). Despite extensive scholarly attention to the antecedents of comparability, research has scarcely focused on its determinants (Dhole et al. 2015; Francis, Pinnuck, and Watanabe 2014; Imhof, Seavey, and Watanabe 2018). We add to this line of research by studying how TMT characteristics (CEOs’ foreign experiences) explain comparability.

Drawing from the upper echelons’ perspective (Hambrick and Mason 1984) and linking it with agency theory (Jensen and Meckling 1976), we examine how FCEOs affect comparability. Going beyond the traditional demographic factors (e.g., tenure, age, and education), we focus on a distinctive executive attribute: foreign experience. Firms recruiting FCEOs are tempted to operate farther to overseas markets and enhance lucrative advantage of their firms by offering social and human capital. Indeed, the value of foreign experience lies in its irreplaceability and uniqueness (Barney 1991). Additionally, firms led by executives with foreign experience dispense more relevant information, which improves governance mechanisms, lowers agency conflicts, and boosts corporate reputation (Giannetti, Liao, and Yu 2015; Iliev and Roth 2018).

We test our hypotheses in the context of China for two reasons. First, although China’s economy has seen significant growth to evolve as the world’s second largest economy, it is still deemed an emerging economy with underdeveloped corporate governance. Since the 1990s, the government has displayed an intensified attempt to hire foreign talents (Giannetti, Liao, and Yu 2015). In 2002, the government disclosed its National Talent Plan for 2002–2005; for the first time, it put forward talent promotion as a national strategy for advancing the country.² Moreover, “The Thousand Talents Program” in 2008³was introduced to attract foreign experienced talents which offered remarkable benefits (i.e., high salaries, start-up research funds, housing allowances, jobs for spouses, and school for children). These policies have attracted many foreigners to join various organizations in China. Second, despite the Chinese government’s efforts to draw foreign talents, overseas experienced executives working in Chinese firms are still fewer than the demand (Giannetti, Liao, and Yu 2015). Nevertheless, many foreign experienced individuals remain in the TMT (particularly on the position of board chair and CEO) and play important roles in corporate decisions. In sum, the two reasons are sufficiently convincing to test our core hypotheses in the Chinese context.

The CEO post is one of the most important positions. Additionally, it distinguishes itself from other managers in several respects, as it oversees the firm’s operations. In other words, the CEO owns organizational power and decision-making authority (Herrmann and Datta 2005). Thus, we exclusively focus on the position of the CEO when examining the effects of TMT characteristics on comparability. We especially place our research attention on CEOs with foreign professional and/or academic experiences owing to their expert knowledge and skills.

Our results suggest that FCEOs positively affect comparability in a sample of Chinese listed firms from 2005 to 2018. This effect is more pronounced when FCEOs have accounting and financial backgrounds and when they have overseas professional experiences but weaker in high economic policy uncertainty (EPU). Upon further scrutiny, we find that financial reporting quality is an important channel that explains the relationship between FCEOs and comparability. Several additional analyses and robustness checks support these conclusions.

This research contributes to two streams of literature. First, the findings contribute to the upper echelon literature by establishing the possible remarkable influence of TMT characteristics on corporate outcomes. As essential features of the TMT, foreign experiences have gained considerable interest in recent years (Conyon et al. 2019; Iliev and Roth 2018). Our unique contribution to this growing literature lies in examining the relationship between FCEO and comparability. We also enrich the literature by arguing that this relationship is prevalent when CEOs have overseas employment experiences and have an accounting and financial background but weaker with a higher EPU.

Second, our study adds to the comparability literature. In this stream of literature, although substantial attention has focused on how comparability affects corporate outcomes, it scarcely determined the factors affecting comparability, such as uncertainty (Dhole et al. 2021), CSR (Wang, Zhang, and Xu 2020), competition (Imhof, Seavey, and Watanabe 2018), audit committee (Endrawes et al. 2018), and auditor style (Francis, Pinnuck, and Watanabe 2014). We proposed another important factor to this stream of literature by arguing that CEOs foreign experiences can affect comparability. Accordingly, our results further develop the growing literature on factors that influence comparability.

The remainder of this paper is organized as follows. Section 2 sets the background and develops the hypotheses. Section 3 details the methods. Section 4 explains the findings and further analyses. Lastly, Section 5 concludes this study.

2. Background and Hypothesis Development

2.1. Foreign Experienced Executives

Over the last few decades, globalization has caused a profound enhancement in international talent mobility, which may be judged as a new medium of foreign knowledge and skill spillover (Liu et al. 2009). With their superior technical and managerial skills (abilities) acquired abroad, foreign talents employed or pursuing academic qualifications in developed economies have earned tremendous value to their home countries, especially in emerging economies (Dai, Kong, and Liu 2018; Yuan and Wen 2018). Prior research (e.g., Wang et al. 2015) has claimed that overseas experienced talents add value to businesses by stimulating knowledge dissemination, cultivating innovation, raising efficiency, and supporting economic progress of their home countries.

The recent research has substantially focused on the relevance of foreign experienced CEOs for corporate outcomes. Upper echelon theory (UET) states that managerial attributes affect strategic decisions and organizational outcomes (Hambrick and Mason 1984). Among the numerous attributes (e.g., expertise, tenure, education, and gender), foreign experiences play a crucial role in boosting organizational innovation ability and access to larger social networks. Lee and Roberts (2015) found that local companies recruit foreign experienced managers because such experiences can be a vital channel for international knowledge transfer (Liu et al. 2009). Harvey (2009) argued that managers with foreign experiences spur economic progress in emerging economies. Yuan and Wen (2018) established a link between managers’ foreign experiences with increased corporate innovation. Iliev and Roth (2018) argued that foreign experienced managers are associated with better corporate governance practices.

Moreover, Liao, Ma, and Yu (2015) suggested that foreign experienced managers can assist in improving corporate transparency in emerging markets, consequently leading to reduced equity financing. Duan, Hou, and Rees (2020) underscored the potential of hiring foreign experienced directors to help minimize information asymmetry. Recognizing the importance of CEOs’ foreign experiences, we add to the literature on how such experience is related to comparability.

2.2. Financial Statement Comparability

Comparability is a vital and distinguishing attribute of financial reporting. Moreover, the value of accounting information increases when it is comparable. As a qualitative aspect of financial reporting, comparability boosts information quality while minimizing information uncertainty and asymmetry. De Franco, Kothari, and Verdi (2011) created a firm-level measure of comparability, arguing that “the accounting system is a mapping from economic events to financial statements.” In particular, two firms producing similar financial statements have comparable accounting systems for a given set of economic events.

Researchers have scrutinized the relevance of comparable accounting information. To illustrate, De Franco, Kothari, and Verdi (2011) found that comparability boosts the quality of information accessible to analysts, thereby leading to increased analyst following, better forecast accuracy, and less dispersion in earnings forecasts. Research has likewise established the association between comparability and an efficient internal capital market (Cheng and Wu 2018), reduced credit risk uncertainty (Kim, Kraft, and Ryan 2013), higher quality of information (Zhang et al. 2013), enhanced forecasting of future stock prices (Shane, Smith, and Zhang 2014), lower contracting costs in the private loan market (Fang et al. 2016), and improved efficiency in governance mechanisms that restrict management discretion on reporting quality (Sohn 2016).

Although research has largely investigated how comparability affects corporate outcomes, only few have explored the factors that affect comparability. For example, studies have argued that comparability can be determined by mandatory IFRS (Barth et al. 2012), SEC’s XBRL mandate (Dhole et al. 2015), auditor style (Francis, Pinnuck, and Watanabe 2014), proprietary cost of disclosure (Imhof, Seavey, and Smith 2017), and EPU (Dhole et al. 2021). However, evidence on how FCEOs affect comparability is unavailable. Thus, the current research aims to fill this gap.

2.3. FCEOs and Comparability

We argue that FCEOs can increase comparability at least for two reasons. Our first argument is based on the “governance view” which proposes foreign experiences of the CEO serves as a governance mechanism that lessens agency issues, which boosts the comparability of financial statements. In this view, FCEOs align management and shareholder interests and thus diminish agency issues (Giannetti, Liao, and Yu 2015). FCEOs provide addittional information to shareholders, consequently facilitating the effective monitoring of managerial activities (Iliev and Roth 2018; Wen, Cui, and Ke 2020). Studies have presented evidence that firms headed by foreign experienced directors are likely to have lower agency issues, supporting the governance role of foreign experienced directors (Dai, Chao, and Wang 2021; Giannetti, Liao, and Yu 2015). Considering that agency issues have been deemed one of the important causes of lowered information quality and in turn comparability (Kim, Kraft, and Ryan 2013; Sohn 2016), we assume if FCEOs can reduce agency issues, it can significantly increase comparability.

Our second argument originates from the “information quality” channel. Specifically, we assume that FCEOs have a positive link with comparability as foreign experiences boost the financial reporting quality. Given that majority of the CEOs accumulate foreign experiences in developed countries, firms headed by FCEOs are likely to implement effective governance practices (Iliev and Roth 2018). Firms with foreign experienced directors tend to disseminate additional information with lower intensity to hoard negative information, which improves information quality provided by such firms (Cao, Sun, and Yuan 2019). Hence, companies led by FCEOs have lower earnings management, which results in better reporting quality (Dai, Chao, and Wang 2021). Considering that comparability increases with higher reporting quality, FCEOs are expected to enhance comparability because of the higher reporting quality.

Following the preceding arguments, FCEOs act as a governance mechanism that balances management’s interests with those of the shareholders (i.e., disciplining management) and thus enhances the corporate information environment. We present the following argument to support our assumptions. If FCEOs improve the corporate information environment, then comparability, as a qualitative component that enhances the financial reporting environment (by minimizing information asymmetry), should also improve. In this study, we address comparability as a distinct aspect of reporting quality. To the extent that FCEOs possess the cognitive ability to filter important information and better monitoring of management, we predict a positive nexus between FCEOs and comparability. Thus, we propose the following hypothesis:

Hypothesis 1. Firms with FCEOs have more comparable financial statements than those led by CEOs without foreign experiences.

2.4. FCEOs’ Accounting and Financial Expertise and Comparability

On the basis of UET, the current study predicts a positive relationship between FCEOs with accounting and financial background and comparability for the following reasons. First, an accumulation of professional accounting and financial expertise through foreign experience grants FCEOs a competitive edge (advantage), which enhances comparability. High-quality disclosure by FCEOs may diminish information asymmetry while boosting governance mechanisms, leading to a better information environment and motivating potential stakeholders (capital market participants) to seek additional firm-specific information. Second, with an enhanced appreciation on how foreign organizations operate, FCEOs with financial expertise are better equipped to execute outstanding management practices and supervise financial policy implementation; resulting in enhanced firm-level governance and improved accounting quality (Dai, Chao, and Wang 2021; Giannetti, Liao, and Yu 2015). Finally, FCEOs often acquire their accounting and financial experiences from developed countries with well-protected legal and commercial background. Accordingly, they display higher ethical standards, which then lead to strong corporate governance. Overall, recruiting FCEOs with accounting and financial expertise may minimize information asymmetry and agency issues and result in improved comparability. Thus, we propose the following hypothesis:

Hypothesis 2. Ceteris paribus, the effect of FCEOs on comparability is more pronounced for FCEOs with accounting and financial backgrounds.

3. Research Design

3.1. Data and Sample Selection

All A-share listed firms on the Shanghai and Shenzhen Stock Exchanges from 2005 to 2018 are initially included in our sample. Following the literature (e.g., Dhole et al. 2021; Yuan and Wen 2018), we eliminate financial firms (due to differences in structure), ST and PT firms,⁴ and industries with fewer than 15 firm–year observations. With these criteria implemented, we finally collect a sample of 1,903 unique firms and 922 unique FCEOs, consisting of 15,241 firm–year observations from 2005 to 2018. The China Securities Market and Accounting Research database is used to collect data on comparability, FCEOs, governance variables, and other financial information. All continuous variables are winsorized to the 99th percentile to reduce the influence of outliers.

3.2. Variables

3.2.1. Dependent Variable: Comparability

Comparability can be measured in two ways: input-based and output-based methods. In input-based methods, one firm’s accounting choices are compared with those of another firm. Different studies (e.g., Barth 2013; Fang et al. 2016) confirmed that input-based methods encounter several challenges such as making a decision which accounting method to use, how such methods should be weighed, and how data can be collected for various accounting choices. Fang et al. (2016) argued that less information regarding the fundamentals of individual firms can be obtained from using common methods (i.e., same set of methods or rules) than firm-specific methods. Barth (2013) also suggested “that using the same set of rules, methods, structures, and content by a firm over the period of time is called consistency.” Similarly, uniformity indicates utilizing the same rules, methods, and contents of financial statements by different firms.

Following prior studies (e.g., Dhole et al. 2021; Kim et al. 2020), we utilize output-based (firm specific) measure of comparability as developed by De Franco, Kothari, and Verdi (2011). The way economic events in a particular span are mapped into accounting figures determines the closeness or similarity of accounting systems. To attain comparable financial statements between two firms, their accounting systems are identified as “mapping from economic events (measured by stock returns) into financial statements (earnings).” We utilize quarterly stock returns and quarterly income to estimate the following regression in measuring a firm’s yearly accounting function.

(1) $Earning{s}_{it}=\hspace{0.17em}{\beta }_i+\hspace{0.17em}{\delta }_{i}Retur{n}_{it}+\hspace{0.17em}{\in }_{it},$(1)

where $Earnings$ denotes earnings outcomes, indicating quarterly net income before extraordinary items divided by market value of equity at the start of period. $Return$ denotes quarterly stock which signifies economic events. We estimate the accounting function for firm $i$ and accounting function for firm $k$ for a given fiscal year. When the closeness of the functions increases, the comparability between two firms increases. To estimate the predicted earnings, we compute accounting functions if firms $i$ and $k$ have both experienced the same economic event.

(2) $\mathrm{F}\mathrm{i}\mathrm{r}\mathrm{m}i:E{\left(Earnings\right)}_{iit}={\hat{\alpha }}_i+{\hat{\beta }}_{i}Return{s}_{it}.$(2)

(3) $\mathrm{F}\mathrm{i}\mathrm{r}\mathrm{m}{}_k:E{\left(Earnings\right)}_{ikt}={\hat{\alpha }}_j+{\hat{\beta }}_{k}Return{s}_{it}$(3)

The earnings in the above equations signify the predicted earnings of both firms $i$ and $k$, with the stock return of firm $i$ in period $t$ as the same economic event. The variance in the accounting systems of two firms would be smaller when firms have comparable accounting systems. We estimate comparability using the following equation.

(4) $Com{p}_{ikt}=\left(-\frac{1}{16}\right)\times {\sum }_{t-15}^t|\mathrm{E}\left(Earnin{g}_{iit}\right)-\mathrm{E}\left(Earnin{g}_{ikt}\right)|$(4)

The $Com{p}_{ikt}$ denotes the comparability between firms “i “and “k” calculated by the absolute difference between the earnings predicted in the above equation for firms “i “and “k” which is then multiplied by −1 for ease of explanation. Therefore, the measure of comparability ($Com{p}_{ikt}$) has negative values. Thus, higher values depict greater comparability between firm $i$ and firm $\mathrm{k}.\hspace{0.17em}$ All the values of $Com{p}_{ikt}$ are ranked, within industry, from highest to lowest to estimate the firm-year level of the comparability measures. We use three firm-level measures of comparability $FSC{4}_{it}$, $FSC{10}_{it}$, and $FSCIn{d}_{it}$. $FSC{4}_{it}$ and $FSC{10}_{it}$ denote the mean values of the four and ten largest comparability scores for firm $i$ during year $t$, while $FSCInd$ denotes the median value of industry comparability scores.

3.2.2. Independent Variable: FCEOs

Following Duan, Hou, and Rees (2020), FCEOs are identified as those possessing overseas working and/or study experiences. We exclude all CEOs who worked, studied, or lived in Hong Kong, Macau, and Taiwan. We create a dummy variable (FCEO) that takes the value of 1 if the CEOs have overseas working and study experiences; otherwise, 0.

We also divided FCEO into two variables: FCEOEXP and FCEOEXP_Non. FCEOEXP (FCEOEXP_Non) is set to 1 if the FCEOs have accounting and financial backgrounds (do not have accounting and financial background); otherwise, 0.

3.3. Empirical Model

The following model is estimated to test the relationship between FCEOs and comparability:

(5) $Comparabilit{y}_{it}={\beta }_0+{\beta }_1{\left(FCEO/FCEOEXP\right)}_{it}+{\beta }_{2}Control{s}_{it}+{\gamma }_{it},$(5)

where $Comparabilit{y}_{it}$ refers to financial statement comparability (as defined in Eq. [4](4) $Com{p}_{ikt}=\left(-\frac{1}{16}\right)\times {\sum }_{t-15}^t|\mathrm{E}\left(Earnin{g}_{iit}\right)-\mathrm{E}\left(Earnin{g}_{ikt}\right)|$(4) ). The main independent variables are FCEO and FCEOEXP. In Eq. (5(5) $Comparabilit{y}_{it}={\beta }_0+{\beta }_1{\left(FCEO/FCEOEXP\right)}_{it}+{\beta }_{2}Control{s}_{it}+{\gamma }_{it},$(5) ), $Control{s}_{it}$ denotes the control variables used in our study following the literature. First, we control for firm size (SIZE: natural log of total assets) and market-to-book ratio (MTB: ratio of market-to-book value of equity). Second, we follow Dhole et al. (2021) and Francis, Pinnuck, and Watanabe (2014) and control for leverage (LEV: total liabilities divided by total assets), operating cash flow (OCF: operating cash flow scaled by total assets), cash flows volatility (OCFV: standard deviation of quarterly operating cash flows over the year), sales growth (SALE: yearly sales growth), and sales volatility (SALEV: standard deviation of quarterly sales over the preceding year). Third, we use return volatility (RETV: standard deviation of daily stock returns over the year) to control for operating risks (De Franco, Kothari, and Verdi 2011). Fourth, we use profitability (ROA: net income scaled by total assets) as a control variable, because it can affect accounting comparability (Sohn 2016). Fifth, we control for accrual quality (AQ)⁵ as a measure of information asymmetry (managerial opportunism). Sixth, we control for certain board and CEOs’ characteristics, such as board independence (BIND: proportion of independent directors on board), duality (DUAL: equals 1 if CEO also serves as board chair; otherwise, 0), board size (BSIZE: total number of directors on the board), female CEO (GCEO: equals 1 if the CEO is female; otherwise, 0), and CEOs’ age (AGE: natural logarithm of CEO’s age). Seventh, we control for ownership and audit quality factors, namely, top shareholders (TOP: percentage of shares held by largest shareholders) and big four auditors (BIG4: equals 1 if the auditor is from one of the big four auditing firms; otherwise, 0). Finally, we include industry and year fixed effects. Appendix A presents the detailed definitions of the variables.

4. Results

4.1. Summary Statistics

Panel A of Table 1 reports the frequency distribution of FCEO across the years. The yearly distribution of FCEO indicates an increasing trend over time with only 9 FCEOs in 2005 and 160 in 2018. The statistics indicate that FCEOEXP also increases from 0 in 2005 to 16 in 2018.

Table 1. Distribution of FCEO.

This table presents the sample distribution. Panel A presents the distribution of FCEOs and FCEOEXP. Panels B and C report the univariate analysis between dependent and independent variables, respectively, and Panel D provides the results of summary statistics. See Appendix A for variables definitions.
Year		N		FCEO		FCEOEXP
2005		695		9		0
2006		631		7		0
2007		710		8		0
2008		799		25		2
2009		853		29		5
2010		861		36		6
2011		980		42		7
2012		1,187		48		8
2013		1,168		71		9
2014		1,265		95		10
2015		1,360		120		11
2016		1,464		131		12
2017		1,581		141		13
2018		1,687		160		16
Total		15,241		922		99
Panel B: Univariate analysis of FCEO and comparability
	FCEO = 1			FCEO = 0
Variables	Mean	SD	N	Mean	SD	N	t-test
FSC4	−0.345	0.868	922	−0.534	1.353	14,319	−4.62***
FSC10	−0.494	1.065	922	−0.740	1.592	14,319	−5.11***
FSCInd	−1.258	1.556	922	−1.675	2.079	14,319	−6.58***
Panel C: Univariate analysis of FCEOEXP and comparability
	FCEOEXP = 1			FCEOEXP = 0
Variables	Mean	SD	N	Mean	SD	N	t-test
FSC4	−0.754	2.052	99	−0.520	1.322	15,142	−1.90*
FSC10	−0.991	2.379	99	−0.723	1.558	15,142	−1.84*
FSCInd	−1.909	2.979	99	−1.646	2.045	15,142	−1.89*
Panel D: Descriptive Statistics
Variables	Obs	Mean	SD	25%	Median	75%
FSC4	15241	−0.648	1.474	−0.549	−0.268	−0.143
FSC10	15241	−0.903	1.736	−0.848	−0.433	−0.230
FSCInd	15241	−2.032	2.207	−2.214	−1.492	−1.002
FCEO	15241	0.060	0.219	0.000	0.000	0.000
FCEOEXP	15241	0.006	0.071	0.000	0.000	0.000
FCEOEXP_Non	15241	0.054	0.208	0.000	0.000	0.000
SIZE	15241	22.104	1.219	21.253	21.976	22.827
MTB	15241	2.662	2.031	1.395	1.995	3.111
LEV	15241	0.472	0.202	0.320	0.480	0.625
ROA	15241	0.035	0.057	0.012	0.031	0.059
OCF	15241	0.047	0.074	0.006	0.045	0.088
OCFV	15241	0.052	0.040	0.024	0.040	0.066
SALE	15241	0.305	0.943	−0.052	0.125	0.351
SALEV	15241	1.74e+09	1.22e+10	1.37e+08	3.49e+08	9.46e+08
AQ	15241	0.250	1.375	0.052	0.117	0.231
RETV	15241	−0.004	1.732	−0.663	0.120	0.838
BIND	15241	0.397	0.083	0.333	0.375	0.444
BSIZE	15241	9.410	2.549	8.000	9.000	11.000
GCEO	15241	0.052	0.222	0.000	0.000	0.000
DUAL	15241	0.196	0.397	0.000	0.000	0.000
BIG4	15241	0.060	0.238	0.000	0.000	0.000
AGE	15241	3.875	0.133	3.784	3.891	3.970
TOP	15241	0.191	0.393	0.000	0.000	0.000

Panels B and C of Table 1 compare the sample mean for firms led by FCEOs with those led by local CEOs. In comparability proxies, all differences are statistically significant at the 1% level. Overall, univariate analysis suggests significant statistical differences for firms led by FCEOs with those led by local CEOs.

Panel D of Table 1 presents the summary statistics for the variables utilized in the analysis. The mean values (SD) of FSC4, FSC10, and FSCInd are −0.648 (1.474), −0.903 (1.736), and −2.032 (2.207), respectively. Overall, the statistics of the comparability measures are consistent with those in the Chinese context (e.g., Yang, Lu, and Tan 2021). The average of FCEO is 0.06%, indicating that, on average, Chinese firms have approximately 6% of CEOs with foreign experience. FCEOs with accounting and financial backgrounds account for 0.6% of the total.

Regarding control variables, the firms in our sample have an average firm size of 22.104 (in natural logarithm), MTB of 2.662, LEV of 47.2%, ROA of 3.50%, OCF of 4.7%, OCFV of 5.2%, SALE of 30.5%, AQ of 0.250, and RETV of −0.004. In addition, the mean (SD) values of BIND, BSIZE, GCEO, DUAL, BIG4, AGE, and TOP are 0.397 (0.083), 9.410 (2.549), 0.052 (0.222), 0.196 (0.397), 0.060 (0.238), 3.875 (0.133), and 0.191 (0.393), respectively.

4.2. FCEOs and Comparability

Table 2 presents the results of the main findings. Columns (1)–(3) present the results of Hypothesis 1, whereas columns (4)–(6) show those of Hypothesis 2. Columns (1)–(3) indicate significant positive coefficients on FCEO, confirming that comparability increases with FCEOs. The findings confirm our argument FCEOs is positively associated with comparability, thereby supporting Hypothesis 1. Columns (4)–(6) indicate that coefficients on FCEOEXP are positive and statistically significant in all cases, thereby supporting Hypothesis 2. Moreover, the coefficients on FCEOEXP are larger than FCEOEXP_Non. Thus, FCEOs with accounting and financial expertise further increase comparability than those with only overseas study experiences. The results suggest that FCEOs’ personnel exchanges, cognitive ability, and international accounting and financial expertise make accounting statements more comparable.

Table 2. Foreign experienced CEOs and comparability.

	(1)	(2)	(3)	(4)	(5)	(6)
	Hypothesis 1			Hypothesis 2
Variables	FSC4	FSC10	FSCInd	FSC4	FSC10	FSCInd
FCEOs	0.092**	0.104**	0.113**
	(2.474)	(2.371)	(1.979)
FCEOEXP				0.258***	0.354***	0.686***
				(6.483)	(7.595)	(11.379)
FCEOEXP_Non				0.131***	0.172***	0.279***
				(3.306)	(3.631)	(4.399)
SIZE	0.121***	0.131***	0.122***	0.110***	0.133***	0.197***
	(6.829)	(6.325)	(4.813)	(7.478)	(7.772)	(9.130)
MTB	−0.003***	−0.003***	−0.003***	−0.002***	−0.003***	−0.003***
	(−3.352)	(−3.380)	(−3.366)	(−3.824)	(−3.973)	(−3.741)
LEV	−0.944***	−1.029***	−1.115***	−1.095***	−1.220***	−1.696***
	(−4.541)	(−4.286)	(−3.824)	(−6.026)	(−5.909)	(−6.850)
ROA	−0.025***	−0.030***	−0.032***	−3.793***	−4.342***	−4.676***
	(−2.785)	(−2.833)	(−2.775)	(−10.069)	(−9.881)	(−8.842)
OCF	1.181***	1.235***	1.073***	0.251	0.481*	0.976***
	(5.049)	(4.551)	(3.239)	(1.072)	(1.759)	(2.906)
OCFV	−2.227**	−2.707**	−3.771**	−2.739**	−3.399**	−4.856*
	(−1.986)	(−1.992)	(−1.987)	(−1.996)	(−2.018)	(−1.960)
SALE	−0.044**	−0.046*	−0.051*	−0.060***	−0.066***	−0.081***
	(−2.102)	(−1.941)	(−1.792)	(−3.745)	(−3.623)	(−3.662)
SALEV	−0.000**	−0.000**	−0.000***	−0.000	−0.000	−0.000
	(−2.501)	(−2.260)	(−2.629)	(−1.624)	(−1.592)	(−0.290)
AQ	0.031**	0.036**	0.040**	0.024***	0.027***	0.029***
	(2.381)	(2.371)	(2.103)	(3.609)	(3.652)	(2.937)
RETV	−0.001	−0.002	−0.004	−0.003	−0.003	−0.003
	(−0.193)	(−0.326)	(−0.408)	(−0.507)	(−0.419)	(−0.276)
BIND	0.095	0.123	0.336	−0.075	−0.121	−0.679***
	(0.626)	(0.698)	(1.541)	(−0.560)	(−0.773)	(−3.437)
DUAL	−0.019	−0.025	−0.089*	0.045	0.068**	0.079*
	(−0.579)	(−0.656)	(−1.907)	(1.545)	(1.981)	(1.794)
BSIZE	−0.011**	−0.015**	−0.016**	−0.008	−0.011*	−0.007
	(−2.045)	(−2.337)	(−2.086)	(−1.497)	(−1.930)	(−0.928)
BIG4	−0.228***	−0.275***	−0.390***	−0.325***	−0.398***	−0.599***
	(−4.123)	(−4.265)	(−4.800)	(−5.104)	(−5.387)	(−6.442)
GCEO	−0.097*	−0.096	−0.149*	−0.067	−0.054	−0.104
	(−1.701)	(−1.461)	(−1.846)	(−1.079)	(−0.745)	(−1.135)
AGE	−0.141	−0.171	−0.153	0.260***	0.368***	0.621***
	(−1.457)	(−1.519)	(−1.095)	(2.683)	(3.237)	(4.268)
TOP	−0.024	−0.033	−0.018	−0.159***	−0.222***	−0.331***
	(−0.755)	(−0.880)	(−0.373)	(−4.586)	(−5.368)	(−6.160)
Year	Yes	Yes	Yes	Yes	Yes	Yes
Industry	Yes	Yes	Yes	Yes	Yes	Yes
Constant	−3.397***	−3.967***	−4.850***	−3.961***	−5.087***	−8.299***
	(−5.981)	(−5.924)	(−5.703)	(−7.576)	(−8.138)	(−9.997)
R^2	0.094	0.114	0.154	0.088	0.089	0.088
N	15,241	15,241	15,241	15,241	15,241	15,241

This table presents OLS regression results with robust standard errors clustered at firm. The t-values are reported in parenthesis. ^{***, **}, and *, denote p < 1%, 5%, and 10%. Refer to Appendix A for variable definitions.

The control variables have similar results consistent with literature (e.g., Dhole et al. 2021; Francis, Pinnuck, and Watanabe 2014). Specifically, in column (1), the coefficients are statistically negative on MTB, LEV, ROA, OCFV, SALE, SALEV, BSIZE, BIG4, and GCEO; whereas statistically positive on SIZE, OCF, and AQ.

5. Robustness Checks and Additional Analyses

To further check the validity of the findings, we opt to carry out several robustness and additional analyses. For brevity and to conserve space, we only report the results of main variables. The detailed description can be found in online supplementary files.

5.1. Alternative Measures of Comparability

Following previous research (e.g., Campbell and Yeung 2016; Chircop et al. 2020), we use FSC4_ADJ, FSC10_ADJ, and FSCInd_ADJ as the new comparability proxies. The detailes are available in Online Suppelementary Files. Panel A of Table 3 presents a summary of the findings using these new comparability measures. In all cases, the coefficients on FCEO and FCEOEXP have statistical significance and are positive.

Table 3. Robustness checks.

Panel A: Alternative proxy of comparability
	Hypothesis 1			Hypothesis 2
Variable	FSC4_ADJ	FSC10_ADJ	FSCInd_ADJ	FSC4_ADJ		FSC10_ADJ		FSCInd_ADJ
FCEO	0.113**	0.122**	0.119*
	(2.214)	(2.195)	(1.865)
FCEOEXP				0.125**		0.141**		0.162***
				(2.299)		(2.441)		(2.598)
FCEOEXP_Non				0.101*		0.109*		0.115*
				(1.854)		(1.843)		(1.685)
Controls	Yes	Yes	Yes	Yes		Yes		Yes
Year	Yes	Yes	Yes	Yes		Yes		Yes
Industry	Yes	Yes	Yes	Yes		Yes		Yes
Constant	−4.910***	−5.375***	−6.133***	−3.286***		−3.534***		−4.580***
	(−6.631)	(−6.647)	(−6.578)	(−4.851)		(−4.731)		(−5.070)
R^2	0.109	0.123	0.154	0.116		0.126		0.163
N	15,241	15,241	15,241	15,241		15,241		15,241
Panel B: Alternative proxy of comparability
	Hypothesis 1			Hypothesis 2
Variable	FSC4_Lag	FSC10_Lag	FSCInd_Lag	FSC4_Lag	FSC10_Lag		FSCInd_Lag
FCEO	0.111**	0.122**	0.119*
	(2.217)	(2.207)	(1.829)
FCEOEXP				0.360***	0.451***		0.778***
				(4.275)	(4.622)		(5.584)
FCEOEXP_Non				0.160**	0.197**		0.288**
				(2.673)	(2.767)		(2.858)
Controls	Yes	Yes	Yes	Yes	Yes		Yes
Year	Yes	Yes	Yes	Yes	Yes		Yes
Industry	Yes	Yes	Yes	Yes	Yes		Yes
Constant	−4.989***	−5.489***	−6.330***	−5.597***	−6.630***		−9.845***
	(−6.566)	(−6.493)	(−6.396)	(−5.620)	(−5.146)		(−4.307)
R^2	0.096	0.111	0.149	0.093	0.095		0.094
N	15,241	15,241	15,241	15,241	15,241		15,241
Panel C: Fixed effect model
	Hypothesis 1			Hypothesis 2
Variable	FSC4	FSC10	FSCInd	FSC4	FSC10		FSCInd
FCEO	0.221***	0.232***	0.217**
	(3.487)	(3.179)	(2.419)
FCEOEXP				0.157***	0.152**		0.625***
				(4.572)	(2.228)		(7.442)
FCEOEXP_Non				0.018	0.022		−0.060
				(0.242)	(0.164)		(−0.348)
Controls	Yes	Yes	Yes	Yes	Yes		Yes
Year fixed effect	Yes	Yes	Yes	Yes	Yes		Yes
Firm fixed effect	Yes	Yes	Yes	Yes	Yes		Yes
Constant	−8.510***	−11.585***	−13.115***	−23.937**	−12.482***		−22.810***
	(−13.152)	(−20.438)	(−14.342)	(−2.403)	(−8.888)		(−7.172)
R^2	0.065	0.073	0.144	0.059	0.073		0.127
N	15,241	15,241	15,073	15,241	15,241		15,241
Panel D: Lag of independent variables
	Hypothesis 1			Hypothesis 2
Variable	FSC4	FSC10	FSCInd	FSC4	FSC10		FSCInd
LFCEO	0.094**	0.111**	0.103*
	(2.299)	(2.316)	(1.695)
LFCEOEXP				0.362***	0.473***		0.811***
				(10.488)	(11.482)		(13.718)
LFCEOEXP_Non				0.076*	0.096*		0.140**
				(1.762)	(1.869)		(1.971)
Controls	Yes	Yes	Yes	Yes	Yes		Yes
Year	Yes	Yes	Yes	Yes	Yes		Yes
Industry	Yes	Yes	Yes	Yes	Yes		Yes
Constant	−3.356***	−3.856***	−4.703***	−3.292***	−4.100***		−6.515***
	(−5.827)	(−5.685)	(−5.498)	(−6.861)	(−7.234)		(−9.041)
R^2	0.092	0.113	0.158	0.087	0.088		0.089
N	13,155	13,155	13,155	13,155	13,155		13,155
Panel E: GMM
	Hypothesis 1			Hypothesis 2
Variable	FSC4	FSC10	FSCInd	FSC4	FSC10		FSCInd
FCEO	0.101***	0.097***	0.191***
	(2.993)	(2.925)	(3.24)
FCEOEXP				0.079***	0.127***		0.242***
				(6.257)	(8.227)		(10.233)
FCEOEXP_Non				0.055*	0.066**		0.118**
				(1.917)	(2.104)		(2.078)
LFSC4	0.973***			0.826***
	(55.490)			(58.886)
LFSC10		1.020***			0.807***
		(61.407)			(67.645)
LFSCInd			1.093***				1.006***
			(93.490)				(117.189)
Controls	Yes	Yes	Yes	Yes	Yes		Yes
Year	Yes	Yes	Yes	Yes	Yes		Yes
Industry	Yes	Yes	Yes	Yes	Yes		Yes
Constant	17.363	4.761	−2.930***	−1.430***	−1.832***		−2.359***
	(1.505)	(0.413)	(−6.741)	(−5.364)	(−5.835)		(−6.243)
Diagnostic Tests
Ar(1)	−7.42***	−7.40***	−7.81***	−7.32***	−7.03***		−7.75***
Ar(2)	1.25	0.65	−0.56	1.36	0.66		−0.59
J-Stats	41.45	44.62	61.57	68.12	72.12		74.81
N	13,296	13,296	13,296	13,296	13,296		13,296
Panel F: 2SLS
	Hypothesis 1			Hypothesis 2
Variable	FSC4	FSC10	FSCInd	FSC4	FSC10		FSCInd
FCEO	0.822**	1.090***	1.744***
	(2.463)	(2.781)	(3.392)
FCEOEXP				9.076*	12.530*		14.511*
				(1.737)	(1.733)		(1.919)
FCEOEXP_Non				0.084**	0.099**		0.101
				(2.077)	(2.041)		(1.524)
Controls	Yes	Yes	Yes	Yes	Yes		Yes
Year	Yes	Yes	Yes	Yes	Yes		Yes
Industry	Yes	Yes	Yes	Yes	Yes		Yes
	−3.594***	−4.232***	−5.289***	−2.627**	−3.262*		−6.638***
	(−6.255)	(−6.255)	(−6.180)	(−2.061)	(−1.889)		(−3.473)
R^2	0.094	0.114	0.154	0.045	0.047		0.053
Wald Chi^2/ F	4061.86	3952.30	4033.89	4249.08	3805.50		3764.86
Prob > F	0.000	0.000	0.000	0.000	0.000		0.000
N	15,241	15,241	15,241	15,241	15,241		15,241
Panel G: Propensity score matching method
Variable	Firm-year obs. with FCEOs (N = 776)		Firm-year obs. without FCEOs (N = 776)		Differences		t-stat
FSC4	−0.413		−0.556		0.142***		2.38
FSC10	−0.595		−0.788		0.192***		2.68
FSCInd	−1.571		−1.854		0.283***		3.01

Notes: Panel A and B report the alternative measures of comparability. Panel C presents the results of FEM. Panel D reports the results of lag variables. Panel E reports the results of GMM. Panel F reports the results of 2SLS. And Panel G reports the results of PSM. The t-values are reported in parenthesis. ^{***, **}, and *, denote p < 1%, 5%, and 10%. Refer to Appendix A for variable definitions.

Second, following De Franco, Kothari, and Verdi (2011), we use lagged stock returns in Eq. (2)(2) $\mathrm{F}\mathrm{i}\mathrm{r}\mathrm{m}i:E{\left(Earnings\right)}_{iit}={\hat{\alpha }}_i+{\hat{\beta }}_{i}Return{s}_{it}.$(2) while estimating the comparability measures, such that accounting earnings are related to lagged stock returns. Collins et al. (1994) argued that stock prices include firm-specific news before they are reported in accounting earnings (i.e., “price lead earnings”). We use the lagged price changes into our model and re-estimate comparability as follows:

(6) $Earning{s}_{it}={\beta }_0+{\beta }_{1}Retur{n}_{it}+{\beta }_{2}Retur{n}_{it-1}+{\epsilon }_{it},$(6)

where $Retur{n}_{it-1}$ is the quarterly lagged stock returns. FSC4_Lag, FSC10_Lag, and FSCInd_Lag are the revised firm–year proxies of comparability based on this model. Panel B of Table 3 presents the results obtained using these alternative proxies of comparability, which are consistent and which provide support for the hypotheses.

5.2. Endogeneity Checks

The previous section reflects our finding that FCEO (FCEOEXP) positively affects comparability. However, this effect may suffer from endogeneity issues. To address endogeneity concerns, we present the five additional analyses in the following sections to account for and reduce endogeneity.

First, we use an alternative approach, i.e., fixed effect model (FEM), to control for potential issues resulting from the omission of time-invariant firm-specific attributes. The results reported in Panel C of Table 3 provide consistent evidence in support of our hypotheses. Second, we use the lag of independent variables, namely, LFCEO and LFCEOEXP, to address endogeneity issues. Panel D of Table 3 shows the results indicating that FCEOs and FCEOEXP improve comparability in Chinese firms. Third, we use GMM to resolve possible endogeneity issues. The findings presented in Panel E of Table 3 remain consistent. Fourth, we use an instrumental variable method, namely, 2SLS, to account for any endogeneity issues. Following Wen, Cui, and Ke (2020) and Yuan and Wen (2018), we use the industry mean percentage of appointing FCEOs as an instrumental variable.⁶ The theoretical logic behind the selection of this instrument is that firms in a given industry are likely to follow similar patterns with other firms in the same industry (Karaevli and Zajac 2013). Further, the average number of FCEOs in an industry is related to whether the firm in the same industry has returnee CEOs or not. Additionally, we assume that this instrument cannot affect comparability via another channel rather than obtaining a foreign experienced CEO. Therefore, we believe that the industry mean percentage of FCEOs can be an instrumental variable. We use the endogenous variables (FCEOs and FCEOEXP) as dependent variables in the first stage, and the other variables, including the instrumental variable (i.e., Industry_Mean),⁷ as independent variables. In the second stage, we regress the comparability measures on the fitted values of FCEOs and FCEOEXP estimated from the first stage regression and a set of control variables and industry and year fixed effect. The results reported in Panel F of Table 3 show that FCEO and FCEOEXP increase comparability. Finally, to address endogeneity issues, we use propensity score matching (PSM) to compare comparability for firms with FCEOs with those without FCEOs. The PSM results presented in Panel G of Table 3 provide consistent evidence.

5.3. Effect of EPU

We argue that EPU can negatively moderate the link between FCEOs and comparability, because EPU raises uncertainty regarding future cash flow, thereby increasing difficulty in making predictions. Consequently, the likelihood that two firms under comparable economic situations will have similar accrual estimates decreases. Particularly, EPU increases the uncertainty of accounting estimations. Nagar, Schoenfeld, and Wellman (2019) found that investors’ reaction to earnings surprises diminishes when EPU is high. Hence, in the scenario of higher EPU, investors are perplexed about the information provided by earnings. Jin et al. (2019) also argued that increased uncertainty in predicting accruals adds difficulty for investors to detect earnings management. Consequently, the possibility of earnings management increases. Considering the likelihood of varying levels of earnings management among firms owing to various objectives, the reported earnings of firms facing relatively comparable economic conditions may also vary and lead to a detrimental effect on comparability.

With the above arguments, EPU is anticipated to negatively moderate the relationship between FCEOs and comparability. Following the literature, we utilize the China EPU index developed by Baker, Bloom, and Davis (2016) to measure policy uncertainty.⁸ (Baker, Bloom, and Davis 2016) used a leading English newspaper⁹ and developed the EPU index for China. This index is based on the textual analysis¹⁰ of news articles related to policy in various media. We re-estimate Eq. (6)(6) $Earning{s}_{it}={\beta }_0+{\beta }_{1}Retur{n}_{it}+{\beta }_{2}Retur{n}_{it-1}+{\epsilon }_{it},$(6) by introducing the interaction of EPU with FCEOs (FCEO*EPU), FCEOEXP (FCEOEXP*EPU) and FCEOEXP_Non (FCEOEXP_Non*EPU). The results in Panel A of Table 4 provide consistent evidence by demonstrating significantly negative coefficients on interaction variables, thereby supporting our arguments.

Table 4. Additional analysis.

Panel A: Moderating effect of EPU
	Hypothesis 1				Hypothesis 2
Variable	FSC4	FSC10	FSCInd		FSC4	FSC10	FSCInd
FCEO	0.228***	0.293***	0.405***
	(5.069)	(4.9714)	(3.5109)
FCEOEXP					0.310***	0.443***	0.827***
					(4.481)	(4.993)	(5.245)
FCEOEXP_Non					0.211***	0.268***	0.364***
					(4.315)	(4.321)	(2.850)
EPU	−0.001***	−0.001***	−0.002***		−0.001***	−0.001***	−0.002***
	(−5.622)	(−5.647)	(−7.085)		(−5.051)	(−5.128)	(−20.013)
FCEO*EPU	−0.000**	−0.000**	−0.001*
	(−3.008)	(−2.799)	(−1.945)
FCEOEXP*EPU					−0.000**	−0.000**	−0.001***
					(−2.471)	(−2.182)	(−3.569)
FCEOEXP_Non*EPU					−0.000*	−0.000**	−0.000
					(−2.015)	(−2.344)	(−1.290)
Controls	Yes	Yes	Yes		Yes	Yes	Yes
Year	Yes	Yes	Yes		Yes	Yes	Yes
Industry	Yes	Yes	Yes		Yes	Yes	Yes
Constant	−4.185***	−5.354***	−8.986***		−4.170***	−5.330***	−8.268***
	(−9.528)	(−8.549)	(−6.651)		(−9.719)	(−8.722)	(−6.795)
R^2	0.0489	0.0538	0.1095		0.049	0.055	0.077
N	15,241	15,241	15,241		15,241	15,241	15,241
Panel B: CEOs working versus studying experiences
Variable	FSC4	FSC10	FSCInd		FSC4	FSC10	FSCInd
FCEO_Work	0.164***	0.236***	0.410***
	(3.152)	(3.871)	(5.153)
FCEO_Study					0.104	0.157	0.007
					(1.333)	(1.587)	(0.057)
Controls	Yes	Yes	Yes		Yes	Yes	Yes
Year	Yes	Yes	Yes		Yes	Yes	Yes
Industry	Yes	Yes	Yes		Yes	Yes	Yes
Constant	−4.747***	−6.182***	−10.011***		−2.959***	−3.469***	−4.643***
	(−9.384)	(−10.142)	(−12.227)		(−6.042)	(−5.963)	(−6.212)
R^2	0.048	0.051	0.057		0.063	0.075	0.15
N	15,241	15,241	15,241		15,241	15,241	15,241
Panel C: Channel Analysis
	Hypothesis 1				Hypothesis 2
Variable	FSC4	FRQ	FSC4		FSC4	FRQ	FSC4
RCEOs	0.093**	0.012**	0.104*
	(2.510)	(2.043)	(1.922)
FRQ			0.261***				0.260***
			(2.713)				(2.671)
FCEOEXP					0.105**	0.018**	0.095**
					(2.544)	(2.546)	(2.232)
FCEOEXP_Non					0.081**	0.010*	0.094**
					(1.976)	(1.797)	(2.512)
Controls	Yes	Yes	Yes		Yes	Yes	Yes
Year	Yes	Yes	Yes		Yes	Yes	Yes
Industry	Yes	Yes	Yes		Yes	Yes	Yes
Constant	−3.373***	0.543	−3.234***		−3.381***	−0.046	−3.241***
	(−5.931)	(0.520)	(−3.880)		(−5.940)	(−0.379)	(−5.962)
R^2	0.093	0.050	0.093		0.093	0.065	0.094
N	15,241	15,046	15,046		15,241	15,046	15,046
Sobel test (p_value)			1.656 (0.099)				1.852(0.064)

Notes: Panel A reports the results of the moderating effect of economic policy uncertainty on the relationship between returnee CEOs and comparability. Panel B reports the results of the effect returnee CEOs with working and studying experiences on comparability. Panel C reports the mediating role financial reporting quality on comparability. The t-values are reported in parenthesis. ^{***, **}, and *, denote p < 1%, 5%, and 10%. Refer to Appendix A for variable definitions.

5.4. CEOs’ Foreign Working versus Study Experiences

CEOs can gain foreign experiences by studying and working abroad. We further divide the sample in two on the basis of CEOs’ foreign experiences, namely, working experiences (FCEO_Work) and studying experiences (FCEO_Study). Panel B of Table 4 reports the regression results. The coefficients on FCEO_Work are significantly positive, whereas those on FCEO_Study are insignificant. This result is consistent with the view that foreign working experiences are deemed fairly important experiences in terms of comparability.

5.5. Channel Analysis

In the previous sections, we have determined that FCEOs can increase comparability. However, we ask the following question: what is the channel through which FCEOs can improve comparability? We respond by arguing that financial reporting quality can explain the effect of FCEOs on comparability. FCEOs have been demonstrated to significantly increase the quality of financial reports by reducing earnings management (Dai, Chao, and Wang 2021). Given that reporting quality is one of the important factors that can improve comparability, FCEOs increase comparability by improving reporting quality.

We follow Kothari, Leone, and Wasley (2005) for financial reporting quality (FRQ) measurement to examine the reporting quality channel. The literature presents several definitions related to the quality of accounting earnings, because it is a multidimensional concept. According to Francis et al. (2005), two groups of earnings attributes are commonly identified to measure financial reporting quality (FRQ)” accounting based attributes (i.e., accrual quality) and the market-based attributes (i.e., timelines, conservatism, and value relevance). We use accounting-based attribute by focusing on the accrual quality as a measure of FRQ, because it has more advantage over other methods. Accrual quality represents an attribute for FRQ that focus on the precision of accounting information. The underline argument behind using accruals quality as a proxy of FRQ is based on the idea that it conveys information about the firm’s operations. Therefore, the accrual component of earnings is a useful measure for FRQ (Dechow, Sloan, and Sweeney 1995). According to the literature, the primary role of accruals is to overcome problems with measuring financial performance, which could be manipulated. The evidence indicates that the probability of manipulation increases with the increase of accruals (Beneish 1999). Therefore, we used Kothari, Leone, and Wasley (2005) model, which is a reliable method for measuring FRQ in literature (Chen et al. 2011). Several studies have employed this method (e.g., Chen et al. 2011), including in the Chinese context (e.g., Jiang and Chen 2019; Trinh, Haddad, and Tran 2022). We use the following performance-adjusted discretionary accruals model following Kothari, Leone, and Wasley (2005) as follows:

(7) $T{A}_{it}={\delta }_0+{\delta }_1\left({\frac{1}{Assets}}_{it-1}\right)+{\delta }_2\mathrm{\Delta }Sale{s}_{it}+{\delta }_{3}PP{E}_{it}+{\delta }_{4}RO{A}_{it}+{\epsilon }_{it},$(7)

where $T{A}_{it}$ represents total accruals, computed as change in non-cash current assets less change in non-interest-bearing liability, less depreciation for firm i at year t, divided by previous year’s total assets. $\mathrm{\Delta }Sale{s}_{it}$ is change in sales divided by previous year’s total assets. $PP{E}_{it}$ denotes net property, plant, and equipment scaled by lagged total assets. $RO{A}_{it}$ is retun on assets for firm i at year t. The residuals obtained from the model are discretionary accruals. The absolute value from the discretionary accruals is proxy used for financial reporting quality (e.g., Chen et al. 2011). For easy interpretation, we multiplied the absolute value of discretionary accruals by −1 (denoted by FRQ). Therefore, higher values represent higher FRQ.

We adopted the Baron and Kenny (1986) step-wise regression method to test the mediating role of financial reporting quality on the relationship between FCEOs and comparability. Specifically, we employed a three-step procedure. Step 1 examines the effect of main independent variable on the dependent variable. Step 2 studies the influence of the independent variable on mediating variable. Step 3 examines the combined effects of independent and mediating variables. To confirm the presence of a mediation effect, we used Sobel (1982) test. The results reported in Panel C of Table 4 confirm the mediating role of FRQ on the relationship between FCEOs (FCEOEXP) and comparability.

6. Conclusion

This study investigates whether FCEOs and their accounting and financial expertise are associated with comparability. It utilizes a sample of Chinese listed firms from 2005 to 2018 and found evidence that FCEOs significantly increase comparability. Notably, this this influence is more pronounced for CEOs with accounting and financial backgrounds. These conclusions remain valid after addressing endogeneity issues. Our additional analyses also suggest the following: the effect of FCEOs on comparability is stronger when CEOs have overseas employment experiences but weaker in the presence of higher EPU. Finally, we identify financial reporting quality as a channel explaining the nexus between FCEOs and comparability.

This research provides new evidence confirming the importance and influence of CEOs’ foreign experiences on comparability. It underscores the significance of foreign experiences as integral in the acquisition of human and social capital. It also adds to the increasing body of knowledge regarding the economic implications of CEOs’ foreign experiences. Our findings are useful to firms that wish to attract foreign talents by recognizing the prospect of FCEOs to minimize agency problems and offer potential external resources. For policymakers, acknowledging that FCEOs can effectively help improve comparability might inspire them to allocate potential resources to crafting policies that attract foreign expertise. Overall, the results offer important implications for policymakers and regulators with regard to the role of FCEOs.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the Research on the development of digitalized rural inclusive financial services systems in Liaoning Province [LN2020Z03]; Research on the bank competition and regional economic growth based on shadow banking channel [20210007].

Notes

1. We use the term CEOs’ foreign experience, FCEOs, and foreign experienced CEOs interchangeably hereafter.

2. See “http://www.moe.gov.cn/jyb_xxgk/gk_gbgg/moe_0/moe_8/moe_26/tnull_404.html” (in Chinese) for more details.

3. See “http://www.1000plan.org/en/” for more details on the Thousand Talents Program.

4. “ST” and “PT” represent special treatment and particular treatment firms. The shares of such firms are omitted due to certain financial issues.

5. AQ means “Absolute value of discretionary accruals” using the Dechow and Dichev (2002) model.

6. We run multiple post-estimation methods, such as F-statistics and t-statistics to check the validity of instrumental variables. The values of F-statistics and t-statistics are higher than the cutoff points suggested by Staiger and Stock (1997) and Adkins and Hill (2008), respectively. Thus, we conclude that our instrument variable is reliable, valid, and adequately strong to minimize endogeneity concerns in our 2SLS regressions.

7. We use industry means for both FCEO and FCEOEXP.

8. “Chinese EPU indexed is developed by finding articles on Chinese EPU for each month and dividing the number of these articles by the total number of articles published that month.”

9. For example, “The South China Morning Post in Hong Kong.”

10. The textual analysis is based on keywords such as “economic/economy,” “uncertain/uncertainty,” “tax,” “policy,” “deficit/budget deficit,” “bank/central bank,” and “regulations.”

Appendix A.

Variable definitions

Table

Variable	Definition
Dependent variables
FSC4	FSC4 is the mean of the four highest comparability scores of firm “$i$” in period “$t$” in the same industry.
FSC10	FSC10 is the mean of the comparability scores of the top 10 firm within the industry.
FSCInd	FSCInd is the median of all of the comparability scores of firm “$i$” in the industry for period “$t$”.
Independent variables
FCEO	Equals 1 if the CEO has foreign working or studying experiences, otherwise 0.
FCEOEXP	Equals 1 if the FCEO has accounting and financial experience, otherwise 0.
FCEOEXP_Non	Equals 1 for the FCEO with no accounting and financial experience, otherwise 0.
Control variables
SIZE	Natural log of total assets.
MTB	Market to book value of equity.
LEV	Total liability over total assets.
ROA	Net income scaled by total assets.
OCF	Operating cash flows deflated by total assets.
OCFV	Standard deviation of quarterly operating cash flows over the preceding year.
SALE	Annual sales growth
SALEV	Standard deviation of quarterly sales over the preceding year.
AQ	Absolute value of discretionary accruals calculated using the Dechow and Dichev (2002) model.
RETV	Standard deviation of daily stock returns over the year.
BIND	Percentage of independent directors
DUAL	Equals to 1 if the CEOs also serves as board chair, otherwise 0.
BSIZE	Total number of board members.
BIG4	Equals 1 if the auditor is from one of the Big 4 firms (auditing), otherwise 0.
GCEO	Equals 1 if the CEO is female, otherwise 0.
AGE	Natural log of CEOs age.
TOP	Percentage of shares held by largest shareholders.
Variables used in additional analyses
EPU	News-based index: EPU index is developed by Baker, Bloom, and Davis (2016). It is calculated on monthly basis. For annual values, the monthly data is converted data by giving the same weight to every month, and then deflated by 100.
FCEO_Work	Equals 1 if CEO has foreign work experience, otherwise 0.
FCEO_Study	Equals 1 if CEO has foreign study experience, otherwise 0.
FRQ	Absolute value of discretionary accruals multiplied by −1 calculated using the Kothari, Leone, and Wasley (2005)