Assessing the Impact of Analyst Coverage and Political Connections on Audit Opinion Shopping in China's Corporate Sector

doi:10.21203/rs.3.rs-5325102/v1

Download PDF

Article

Assessing the Impact of Analyst Coverage and Political Connections on Audit Opinion Shopping in China's Corporate Sector

https://doi.org/10.21203/rs.3.rs-5325102/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

This study builds on the framework established by Lennox (2002) to assess the likelihood of firms receiving modified audit opinions (MAOs) and explores the relationship between analyst coverage and audit opinion shopping among Chinese-listed firms, with a particular focus on the moderating effects of political connections and political leadership turnover. Leveraging a comprehensive dataset of 37,774 observations from 2004 to 2022, our findings indicate that analyst coverage significantly increases the propensity for audit opinion shopping, especially in firms with strong political ties. Additionally, we conduct an event study to examine the impact of analyst announcements within China’s A-share market. By analyzing reports from top-ranked analysts identified through the "New Fortune Best Analyst" rankings, we observe significant market reactions, suggesting that influential analyst coverage heightens the pressure on firms to secure favorable audit outcomes, particularly during critical financial periods. Furthermore, our analysis of political leadership turnover during the 17th, 18th, and 19th National Congresses reveals that changes in local political leadership, particularly the promotion of local officials, further amplify firms’ tendency to engage in audit opinion shopping. These findings underscore the intricate interplay between market scrutiny, political dynamics, and corporate audit practices, providing new insights into the factors that drive audit opinion shopping in China’s distinctive corporate landscape.

Business and commerce/Business and management

Business and commerce/Finance

Analyst Coverage

Audit Opinion Shopping

Political Connections

Political Leadership Turnover

Event Study

Audit opinion shopping, where clients seek auditors more likely to provide favorable opinions when an incumbent auditor may issue an unfavorable one, is a well-explored topic in audit research. Studies have examined this phenomenon across various countries, focusing on its impact on auditor independence and audit quality, particularly in relation to political connections (Lu, 2006; Bliss et al., 2011; Wahab et al., 2013; Newton et al., 2016; Khan et al., 2016). This study seeks to expand on these findings by exploring opinion shopping within China's unique auditing environment, with a specific focus on the interplay between analyst coverage and political connections.

Analyst coverage increases audit opinion shopping by enhancing corporate visibility and investor scrutiny, prompting firms to seek favorable audit outcomes. The established link between analyst coverage and investor attention suggests that companies under scrutiny aim to project positive financial images (Bonner et al., 2007). This motive is amplified by the correlation between analyst focus and stock price volatility (Kothari et al., 2009). Analyst recommendations significantly influence investor behavior, especially in uncertain markets (Zhang and Wei, 2019; Kliger and Kudryavtsev, 2013), affecting firms' audit choices to meet market expectations (Ghosh and Moon, 2005; Newton, 2019). The significant role of analysts during pivotal financial events, like IPOs, indicates firms' tendency towards opinion shopping to secure favorable market reception (Romain Boissin and Sentis, 2012; Rajan and Servaes, 1997). Thus, increased analyst scrutiny leads companies to seek positive audit outcomes, aligning their financial reporting and auditing practices with market and investor expectations.

Political connections also significantly influence audit practices by shaping auditor-client relationships and impacting audit outcomes. Firms with political ties, particularly in China, are more likely to receive favorable audit opinions, possibly due to perceived lower risk or expectations of regulatory leniency (He et al., 2017; Yang et al., 2023; Tantawy and Moussa, 2023). This trend often involves a preference for smaller, local auditors, which can compromise audit independence and quality due to political influences. The global repercussions of political connections affect audit fees, audit quality, and financial reporting practices. For instance, in Malaysia and Korea, political ties have significantly impacted the auditing landscape (Wahab et al., 2011; Wahab et al., 2014; Cho et al., 2017; Tee et al., 2017).

Our study examines not only the direct impact of analyst coverage on audit opinion shopping but also how political connections influence this relationship. Unlike previous research, such as Lin and Wang (2022), which primarily focused on the role of earnings management as a mediator, we explore how political ties shape firms' tendencies to engage in audit opinion shopping under the scrutiny of analysts. Additionally, we conduct an event study to assess the effects of political leadership turnover during the 17th, 18th, and 19th National Congresses on audit opinion shopping. This analysis offers a nuanced perspective on how changes in local political leadership particularly whether officials are promoted locally or parachuted in affect the likelihood of firms engaging in audit opinion shopping.

Our study, based on 37,774 observations from Chinese-listed companies between 2004 and 2022, uses Lennox (2000)’s 'what if' framework to assess the likelihood of receiving a modified audit opinion (MAO) as a measure of audit opinion shopping. The results show that analyst tracking intensifies audit opinion shopping, and political connections further amplify this effect.

The contribution of this study is fourfold. First, our research shows that analyst coverage significantly increases the likelihood of audit opinion shopping Lin and Wang (2022), especially in firms with strong political connections. This finding contributes to the existing literature by revealing that analyst scrutiny, rather than deterring opportunistic behavior, can actually prompt firms to seek more favorable audit opinions, highlighting a complex interaction between analyst monitoring and corporate governance. Second, we perform a detailed event study to examine the impact of analyst announcements on audit opinion shopping within China’s A-share market. By focusing on reports from top-ranked analysts identified through the "New Fortune Best Analyst" rankings, we capture significant market reactions. This part of our study demonstrates that influential analyst coverage can increase the pressure on firms to secure favorable audit outcomes, particularly during critical financial periods, thus enriching the literature on how market reactions to analyst reports influence audit practices (Ayres, 2019). Third, we offer a unique analysis of political leadership turnover, centered on the 17th, 18th, and 19th National Congresses. Our findings indicate that changes in local political leadership, particularly the promotion of local officials versus those parachuted in from elsewhere, have a significant effect on audit opinion shopping behavior. This contribution enhances the understanding of how regional political dynamics influence corporate audit practices, providing new insights into the importance of political stability and leadership continuity in corporate decision-making (Tessema and Abou-El-Sood, 2022). Finally, our findings carry important implications for policymakers, market regulators, and investors. We highlight the need for increased oversight of firms that are both politically connected and under heavy analyst scrutiny, as these firms are more likely to engage in audit opinion shopping (Kallias et al., 2022). Our study suggests that regulatory frameworks should address the combined influence of political and market pressures on corporate behavior, to ensure that both audit independence and the quality of financial reporting are upheld in environments where these factors are particularly influential.

2.1 The Influence of Analyst Coverage on Audit Opinion Shopping

Analysts play a crucial role in external monitoring and information dissemination, which helps curb managerial opportunism. By gathering, processing, and presenting information from both public and private sources, analysts provide valuable insights that go beyond what companies publicly disclose. This reduces information asymmetry between managers and investors, facilitating better external monitoring. Research shows that analysts significantly impact priced and investment risks, with their equity risk rating changes having a greater stock price impact than credit rating changes (Lui et al., 2007; Lui et al., 2012). Analysts' risk spread forecasts capture a firm's operational risk and predict changes in its fundamentals, aiding investor decision-making (Chu et al., 2021). Additionally, analysts' reports often include detailed discussions on a firm's performance, management quality, strategies, and competitiveness, offering a comprehensive view of the company's prospects (Asquith et al., 2005).

Given the critical role of analysts in monitoring and information dissemination, their coverage can significantly influence corporate behaviors, including audit opinion shopping. Audit opinion shopping, where companies seek favorable audit opinions to avoid negative outcomes from current auditors, has been a significant topic of regulatory concern and academic interest. Initial studies, such as those by Dopuch et al.(1986), suggested that this practice might be costly due to the negative perception of Modified Audit Opinions (MAOs). However, earlier research in the U.S. found that switching auditors to avoid unfavorable opinions was generally unsuccessful. More recent studies provide a nuanced understanding. For instance, Lennox (2002) demonstrated that UK companies successfully engaged in opinion shopping, with similar patterns observed in non-U.S. contexts like China (Green et al., 2009; Czernkowski et al., 2010). These studies indicate potential opinion shopping but do not conclusively link it to financial or audit outcomes.

In China's rapidly growing market with evolving regulations, analysts' roles are particularly pertinent. Increased analyst coverage enhances corporate visibility and investor scrutiny (Lang et al., 2004; Wiersema and Zhang, 2011), compelling firms to seek favorable audit outcomes to maintain a positive financial image. The correlation between analyst focus and stock price volatility amplifies this motive, as firms aim to align their financial reporting and auditing practices with market and investor expectations. Analyst recommendations significantly influence investor behavior, particularly in uncertain markets, impacting firms' auditing choices to meet market expectations.

Given the enhanced visibility and scrutiny from analysts, firms may be more inclined to engage in audit opinion shopping to avoid negative audit outcomes. This behavior aims to present a more favorable financial position to satisfy analysts and investors, potentially compromising audit quality.

Therefore, we hypothesize that:

H1: Increased analyst coverage is positively associated with the likelihood of firms engaging in audit opinion shopping.

2.2 The Moderating Role of Political Connections in Audit Opinion Shopping

Political connections can significantly influence audit practices and outcomes. Research indicates that firms with political connections may leverage these ties to secure more favorable audit opinions, potentially at the expense of audit quality. For instance, Gul et al. (2007) found that non-Big 4/5 auditors with political connections are associated with lower audit quality compared to their non-connected counterparts. This tendency is particularly pronounced for economically significant clients, suggesting that political connections may facilitate more lenient audit outcomes.

Additionally, institutional investors' willingness to pay higher audit fees for politically connected firms (Tee et al., 2017) reflects a demand for greater audit effort as a monitoring mechanism. This is especially true among foreign institutions, highlighting global concerns over the influence of political ties on corporate governance. These connections can create an environment where firms are more likely to engage in audit opinion shopping to meet the expectations of analysts and investors.

Therefore, we hypothesize that:

H2a: Political connections strengthen the relationship between analyst coverage and an increased propensity for audit opinion shopping.

Conversely, political connections might also mitigate the pressure to engage in audit opinion shopping. Firms with strong political ties often enjoy regulatory forbearance and reduced scrutiny, which can lead to more stable audit outcomes. He et al. (2017) demonstrated that state-owned enterprises (SOEs) tend to favor local small auditors post-connection, indicating a preference for less stringent audit practices due to their political backing.

Furthermore, the termination of political connections has been shown to influence audit opinions and auditor choice, with non-SOEs favoring larger auditors after losing their political ties(He et al., 2017). This shift suggests that political connections provide a buffer against the need for aggressive audit opinion shopping. In such cases, the role of analyst coverage in prompting firms to seek favorable audit outcomes might be diminished, as political connections already offer a degree of protection.

Therefore, we hypothesize that:

H2b: Political connections weaken the relationship between analyst coverage and an increased propensity for audit opinion shopping.

3.1 Sample selection

Our study examines the dynamics of analyst tracking, audit opinion shopping, and political connections using a sample of A-share listed companies on the Shanghai and Shenzhen stock exchanges from 2004 to 2022. Although analyst tracking data has been available since 2002, our dataset begins in 2004 to ensure the availability and reliability of comprehensive data for our selected variables, allowing for a robust longitudinal analysis.

Consistent with the methodology outlined by Lennox (2000), we focused on the top 30% of listed companies most likely to receive unclean audit opinions. This targeted selection helps us understand the nuances and motivations behind audit opinion shopping behaviors in the Chinese market.

Our research methodology involved a detailed sample selection process starting from an initial pool of 53,081 observations. We excluded companies from the financial industry (1,118 observations), those without current audit opinion types (3,975 observations), companies designated as ST (Special Treatment, 2,983 observations), and samples with missing data (7,231 observations). After these exclusions, our final dataset comprised 37,774 observations.

The dataset includes comprehensive information from the CSMAR database, a reliable and respected resource for research on Chinese markets. This information encompasses financial statements, audit opinions, details of auditors, and analyst tracking activities. To further enhance the robustness of our findings, we applied a 1% tailoring approach to our dataset. Detailed sample construction procedures are summarized in Table 1.

Table 1

Sample Selection
Criteria	Observations Excluded	Cumulative Observations Remaining
Initial Pool		53,081
Exclusion: Financial Industry	(1,118)	51,963
Exclusion: No Current Audit Opinion	(3,975)	47,988
Exclusion: Special Treatment (ST) Status	(2,983)	45,005
Exclusion: Missing Data	(7,231)	37,774
Final Dataset	(15,307)	37,774

3.2 Variable description and Analysis models

3.2.1 Audit opinion shopping

We apply the Lennox (2000)’s Model to measure the audit opinion shopping variable.

$$\:\begin{array}{c}{Q}_{i,t}^{qs}={\alpha\:}_{0}+{\alpha\:}_{1}{\text{Switch}}_{i,t}+{\alpha\:}_{2}{Q}_{i,t-1}+{\alpha\:}_{3}{Q}_{i,t-1}\times\:{\text{Switch}}_{i,t}+{\alpha\:}_{4}{X}_{i,t}+{\alpha\:}_{5}{X}_{i,t}\times\:{\text{Switch}}_{i,t} \left(1\right)\end{array}$$

In this model, the dummy variable $\:{Q}_{i,t}^{qs}$ indicates whether a client may have engaged in opinion shopping. A change in audit opinion over two consecutive years can signal such behavior. Specifically, if a firm receives a nonstandard (unclean) audit opinion in year t, $\:{Q}_{i,t}$ is set to 1; otherwise, it is 0. The variable $\:{Q}_{i,t-1}$ captures the audit opinion the firm received in the previous year (t-1) with 1 representing an unclean opinion and 0 otherwise. The dummy variable $\:{\text{Switch}}_{i,t}$ is assigned a value of 1 if the firm changes its auditor in year 𝑡 ; otherwise, it is 0. Control variables $\:{X}_{i,t}$ include operating cash flows (CF), financial leverage (LEV), return on capital (ROC), internal equity ownership (DIRS), and auditor size (BIG4). Fixed effects for Year and Industry are incorporated to account for temporal and industry-specific factors, with $\:{v}_{i,t}$ representing the error term.

$$\:\begin{array}{c}{\text{S}\text{w}\text{i}\text{t}\text{c}\text{h}}_{i,t}={\beta\:}_{0}+\beta\:{\text{O}\text{P}}_{i,t}+{\beta\:}_{2}{X}_{i,t}+Year+Industry+{v}_{i,t} \left(2\right)\end{array}$$

From Model (1), we calculate the probability $\:{\text{P}\text{r}(Q}_{i,t}^{qs})$ of firm 𝑖 receiving a nonstandard audit opinion in year 𝑡. Model (2) is then developed to determine whether the firm engaged in opinion shopping during that year.

According to Davidson et al. (2006), firms may switch auditors to minimize the probability of receiving a nonstandard opinion if they are engaging in opinion shopping. $\:{\text{O}\text{P}}_{i,t}$ quantifies the difference in the likelihood of receiving a nonstandard audit opinion under scenarios of auditor change versus no change. A positive $\:{\text{O}\text{P}}_{i,t}\text{(i.e.,}Pr({Q}_{i,t}^{q1}=1)>Pr({Q}_{i,t}^{q0}=1))$ suggests that switching auditors would increase the probability of receiving a nonstandard opinion, potentially motivating the firm to retain its current auditor. Conversely, a negative $\:{\text{O}\text{P}}_{i,t}$ suggests a reduced likelihood of receiving a nonstandard opinion, making the firm more likely to switch auditors to engage in opinion shopping.

3.2.2 Analyst Coverage

For each fiscal year, we determine analyst coverage by averaging the 12 monthly earnings forecasts from the summary file, which serves as our raw measure of analyst coverage (Analyst). This approach is based on the assumption that most analysts following a firm will provide at least one earnings forecast within the year leading up to the firm's fiscal year-end, with most issuing no more than one forecast per month. We then calculate the natural logarithm of (one plus) this raw measure to derive our primary measure of analyst coverage. Additionally, we define a binary variable, Analyst Presence, which is set to 1 if a firm's analyst coverage exceeds 1, and 0 otherwise.

3.2.3 Political Connections

In our study, the variable PC is used to represent political connections and is defined as a dummy variable. It is assigned a value of 1 if the firm’s CEO or Chairperson currently holds or has previously held a political position, such as an officer in the central or local government, the military, a member of the People’s Congress (CPC), or the People’s Political Consultative Conference (CPPCC). If these conditions are not met, the variable is set to 0. This approach aligns with prior research (Fan et al., 2007), which identifies political connections based on the political backgrounds of key executives.

While PC captures political ties at the executive level, it is important to recognize that this represents only one aspect of political connections. Another widely used proxy is the percentage of government ownership in a firm, known as SOER. This measure, based on the methodologies of Adhikari et al. (2006), serves as an indicator of government support for certain firms. Although direct equity ownership may not account for all the ways the government supports firms, it remains a practical and widely accepted proxy.

3.3 Analysis models

To test Hypothesis 1, which investigates whether analyst tracking influences a company's propensity for opinion shopping, we extend our analysis by incorporating the analyst tracking variable ($\:{Analyst}_{i,t}$) and its interaction with the probability of receiving an unfavorable audit opinion ($\:{{Analyst}_{i,t}\times\:OpnShop}_{i,t}$). This modification leads to the formulation of Eq. (3). The number of analytical institutions tracking a company serves as a proxy for the extent of analyst coverage.

If the coefficient $\:{\theta\:}_{3}$ in Eq. (3) is negative ($\:{\theta\:}_{3}<0$) indicating a negative correlation between the current period auditor change decision ($\:{\text{S}\text{w}\text{i}\text{t}\text{c}\text{h}}_{i,t}$) and the interaction term ($\:{{Analyst}_{i,t}\times\:OpnShop}_{i,t}$), it suggests that analyst tracking reinforces the company's tendency toward opinion shopping. Conversely, if ($\:{\theta\:}_{3}>0$), indicating a positive correlation, it implies that analyst tracking weakens the company's inclination to engage in opinion shopping.

$$\:\begin{array}{c}{\text{S}\text{w}\text{i}\text{t}\text{c}\text{h}}_{i,t}=\:{\theta\:}_{0}+{\theta\:}_{1}{OpnShop}_{i,t}+{\theta\:}_{2}{Analyst}_{i,t}+{\theta\:}_{3}{{Analyst}_{i,t}\times\:OpnShop}_{i,t}+{\theta\:}_{4}{LEV}_{i,t}\\\:\:+{\theta\:}_{5}{ROC}_{i,t}+{\theta\:}_{6}{CF}_{i,t}+{\theta\:}_{7}{DIRS}_{i,t}+{\theta\:}_{8}{LARS}_{i,t}+{\theta\:}_{9}{BIG4}_{i,t}+{\theta\:}_{10}{SIZE}_{i,t}\:\:\\\:\:\:+{\theta\:}_{10}{ROA}_{i,t}+Industry+Year+{ϵ}_{i,t}\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\: \left(3\right)\end{array}$$

Referring to Eq. (4), Hypothesis 2a posits that political connections, as indicated by SOER, enhance the relationship between analyst tracking and audit opinion shopping. This would be reflected in our model as a significant and negative coefficient $\:{\theta\:}_{5}$ for the interaction term $\:{{{SOER\left(PC\right)}_{i,t}\times\:Analyst}_{i,t}\times\:OpnShop}_{i,t}$. A negative $\:{\theta\:}_{5}$ would suggest that firms with political backing are more likely to engage in opinion shopping when under analyst scrutiny, possibly due to the perceived protection or leverage these connections provide. On the other hand, Hypothesis 2b suggests that political connections might weaken the effect of analyst tracking on opinion shopping. In this case, a positive $\:{\theta\:}_{6}\:$coefficient would indicate that political connections reduce or negate the influence of analyst coverage, potentially because politically connected firms have alternative strategies to secure favorable outcomes despite heightened analyst attention.

To test these hypotheses, we extend our model by incorporating the three-way interaction term $\:{{Analyst}_{i,t}\times\:OpnShop}_{i,t}\times\:{SOER\left(PC\right)}_{i,t}$. Similar to our approach for Hypothesis 1, the sign and significance of the $\:{\theta\:}_{5}$ coefficient will reveal the nature of this relationship. A positive $\:{\theta\:}_{5}$ would support Hypothesis 2a, whereas a negative $\:{\theta\:}_{5}$ would lend support to Hypothesis 2b.

$$\:\begin{array}{c}{\text{S}\text{w}\text{i}\text{t}\text{c}\text{h}}_{i,t}=\:{\theta\:}_{0}+{\theta\:}_{1}{OpnShop}_{i,t}+{\theta\:}_{2}{Analyst}_{i,t}+{\theta\:}_{3}{{OpnShop}_{i,t}\times\:Analyst}_{i,t}\\\:\:+{\theta\:}_{4}{{OpnShop}_{i,t}\times\:SOER\left(PC\right)}_{i,t}+{\theta\:}_{5}{{{SOER\left(PC\right)}_{i,t}\times\:Analyst}_{i,t}\times\:OpnShop}_{i,t}\:\\\:+{\theta\:}_{6}{SOER\left(PC\right)}_{i,t}+{\theta\:}_{7}{LEV}_{i,t}+{\theta\:}_{8}{ROC}_{i,t}+{\theta\:}_{9}{CF}_{i,t}+{\theta\:}_{10}{DIRS}_{i,t}+{\theta\:}_{11}{LARS}_{i,t}\\\:\:\:+{\theta\:}_{14}{BIG4}_{i,t}+{\theta\:}_{14}{SIZE}_{i,t}+{\theta\:}_{15}{ROA}_{i,t}+Industry+Year+{ϵ}_{i,t}\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\: \left(4\right)\end{array}$$

Our analytical model incorporates control variables that account for a firm's financial risk, operational efficiency, governance structure, and auditor reputation, all of which are known to influence audit practices. Financial risk is captured through Leverage (LEV), which affects auditors' risk assessments. Return on Capital (ROC) and Cash Flows (CF) from operations are used to gauge a company's economic health and liquidity, both of which play a role in auditors' evaluations of going concern risks. To address ownership structure, we include the proportion of director-shareholders (DIRS) and the presence of large shareholders (LARS), as these factors can influence company policies and the audit process. The involvement of a Big Four auditor (BIG4) is considered to reflect the impact of auditor quality on the likelihood of opinion shopping. Additionally, company size (SIZE) and Return on Assets (ROA) are included due to their relevance in audit complexity, public scrutiny, and the auditors' assessment of financial performance and stability.

4.1 Descriptive statistics

Table 2 provides the descriptive statistics for the control variables used in this study, offering a detailed view of audit practices among Chinese publicly listed companies. The mean value of Switch is 0.128, indicating that auditor changes are relatively uncommon within the sample, which supports the use of the Lennox (2000) model in our analysis. This infrequency suggests that most companies retain their current auditors, potentially signaling successful opinion shopping among a smaller portion of firms.

The mean value of MAO is 0.054, highlighting the rarity of unclean audit opinions. Financial leverage (LEV) has a mean of 0.462, suggesting that the firms analyzed maintain moderate levels of leverage. The mean Return on Capital (ROC) is 0.231, and the mean Cash Flows (CF) from operations is 0.047, both reflecting generally positive economic health and operational efficiency.

Regarding ownership structure, DIRS and LARS have mean values of 0.079 and 0.409, respectively, indicating the distribution of shareholder influence across firms. The BIG4 variable has a mean of 0.060, implying that only a minority of companies are audited by Big Four firms, which is consistent with the overall landscape of China's audit market. The average firm size (SIZE) in the sample is 22.122, illustrating a broad range of company sizes, while the average Return on Assets (ROA) is 0.630, demonstrating overall profitability.

Moreover, the Analyst variable has a mean of 1.284, showing that, on average, companies are followed by a moderate number of analysts. The Analyst Presence variable, with a mean of 0.689, indicates that the majority of firms have some level of analyst coverage. The political connections variable (PC) shows a mean of 0.318, revealing that roughly one-third of the companies have significant political ties. The SOER variable, which represents state ownership, has a mean of 20.319, showing considerable variation in government ownership levels across the sample.

Table 2

Descriptive statistics
Variables	Count	Mean	SD	Min	P50	P75	Max
Switchi,t	37,774	0.128	0.334	0.000	0.000	0.000	1.000
Analysti,t	37,774	1.284	1.181	0.000	1.099	2.303	3.761
Analyst Presencei,t	37,774	0.689	0.463	0.000	1.000	1.000	1.000
PCi,t	37,774	0.318	0.466	0.000	0.000	1.000	1.000
MAOi,t	37,774	0.054	0.227	0.000	0.000	0.000	1.000
MAOLagi,t	37,774	0.050	0.218	0.000	0.000	0.000	1.000
SOERi,t	37,774	20.319	23.879	0.000	5.315	40.780	77.950
LEVi,t	37,774	0.462	0.218	0.060	0.457	0.614	1.117
ROCi,t	37,774	0.231	0.532	-1.330	0.119	0.273	3.257
CFi,t	37,774	0.047	0.074	-0.186	0.046	0.089	0.257
DIRSi,t	37,774	0.079	0.149	0.000	0.000	0.075	0.613
LARSi,t	37,774	0.409	0.192	0.054	0.411	0.550	0.845
BIG4i,t	37,774	0.060	0.237	0.000	0.000	0.000	1.000
SIZEi,t	37,774	22.122	1.325	19.262	21.966	22.889	26.164
ROAi,t	37,774	0.630	0.451	0.022	0.526	0.780	3.073
NOTE: The table presents the descriptive statistics for the variables applied in the regression analyses. All continuous variables have been winsorized at the 1st and 99th percentiles. For definitions of the variables, refer to the Appendix.

Table 3 presents the Pearson correlations for key variables in our study, revealing a variety of relationship strengths. For example, there is a slight negative correlation between Analyst and Switch (-0.040***), indicating that higher analyst coverage is associated with fewer auditor changes. Additionally, a significant positive correlation exists between MAO and Switch (0.074***), suggesting that receiving a modified audit opinion increases the likelihood of auditor switching. This pattern is similarly observed for the prior year's audit opinions (MAOLag), which shows a correlation of (0.093***) with Switch.

The correlation between financial leverage (LEV) and Switch is positive (0.018***), indicating that firms with higher leverage are more likely to change auditors. Conversely, Return on Capital (ROC) and Cash Flows (CF) show very weak correlations with Switch, at -0.003 and 0.012**, respectively, implying a minimal relationship between financial health and auditor changes.

In terms of ownership structure, DIRS shows a negative correlation (-0.017***) with Switch, while LARS is positively correlated (0.027***), suggesting that companies with concentrated ownership are more likely to switch auditors. The BIG4 variable shows a weak positive correlation (0.010*) with Switch, and firm size (SIZE) is negatively correlated with Switch (-0.032***), indicating that larger companies are less likely to switch auditors. Return on Assets (ROA) shows a very weak positive correlation with Switch (0.006), indicating a minimal relationship. Most correlations are relatively low, with the highest correlations not exceeding 0.591, suggesting that multicollinearity is not a concern in our analysis.

Table 3

Pearson correlation results
	Switch_i,t	MAO_i,t	MAOLag_i,t	Analyst_i,t	Analyst Presence_i,t	PC_i,t	LEV_i,t	ROC_i,t	CF_i,t	DIRS_i,t	LARS_i,t	BIG4_i,t	SIZE_i,t	SOER_i,t	ROA_i,t
Switch_i,t	1
MAO_i,t	0.074^***	1
MAOLag_i,t	0.093^***	0.591^***	1
Analyst_i,t	-0.040^***	-0.189^***	-0.194^***	1
Analyst Presence_i,t	-0.032^***	-0.206^***	-0.217^***	0.770^***	1
PC_i,t	-0.025^***	-0.035^***	-0.033^***	0.087^***	0.084^***	1
LEV_i,t	0.018^***	0.134^***	0.137^***	-0.038^***	-0.044^***	-0.016^***	1
ROC_i,t	-0.003	-0.007	-0.006	0.010^*	0.009	0.010^*	0.026^***	1
CF_i,t	0.012^**	-0.029^***	-0.001	0.040^***	0.024^***	0	0.048^***	0.023^***	1
DIRS_i,t	-0.017^***	-0.038^***	-0.042^***	0.055^***	0.047^***	-0.008	-0.032^***	0.003	0.001	1
LARS_i,t	0.027^***	-0.068^***	-0.054^***	0.103^***	0.090^***	0.012^**	0.009	0.001	0.016^***	-0.269^***	1
BIG4_i,t	0.010^*	-0.037^***	-0.041^***	0.189^***	0.113^***	0.011^*	0.008	0	0.016^***	-0.044^***	0.212^***	1
SIZE_i,t	-0.032^***	-0.135^***	-0.160^***	0.402^***	0.298^***	0.033^***	-0.027^***	-0.007	-0.011^*	-0.112^***	0.306^***	0.340^***	1
SOER_i,t	0.048^***	-0.078^***	-0.060^***	0.045^***	0.051^***	-0.035^***	0.023^***	-0.006	0.010^*	-0.223^***	0.461^***	0.132^***	0.346^***	1
ROA_i,t	0.006	-0.021^***	-0.010^*	0.076^***	0.048^***	-0.027^***	0.019^***	-0.006	0.023^***	-0.020^***	0.064^***	0.028^***	0.022^***	0.047^***	1
* p < 0.1, p < 0.05, * p < 0.01

4.2. Results from the audit reporting model

Table 4 presents the probit regression results from our audit reporting analysis, with a particular focus on the interaction effects shown in Column (2). A key finding is the significant negative coefficient of the interaction between auditor switching and past modified audit opinions (Switchi,t×MAOLagi,t = -0.100***), indicating that companies with a history of MAOs are less likely to receive another MAOi,t when they switch auditors. This suggests that firms may strategically switch auditors to avoid unfavorable audit outcomes, and this behavior is more pronounced when past audit opinions have been modified.

Table 4

Audit reporting model
Variables	(1)	(2)
Variables	MAOi,t	MAOi,t
MAOLagi,t	0.607^***	0.627^***
MAOLagi,t	(121.063)	(107.638)
Switchi,t		0.285^***
Switchi,t		(5.466)
Switchi,t×MAOLagi,t		-0.100^***
Switchi,t×MAOLagi,t		(-8.704)
LEVi,t	0.010^***	0.009^***
LEVi,t	(11.874)	(9.708)
ROCi,t	-0.038	-0.031
ROCi,t	(-0.912)	(-0.309)
CFi,t	-0.018^***	-0.063^***
CFi,t	(-6.606)	(-6.292)
DIRSi,t	-0.021^***	-0.016^***
DIRSi,t	(-5.594)	(-3.914)
LARSi,t	-0.033^***	-0.022^***
LARSi,t	(-5.704)	(-3.580)
BIG4i,t	0.006	0.004
BIG4i,t	(1.315)	(0.729)
SIZEi,t	-0.007^***	-0.005^***
SIZEi,t	(-8.015)	(-5.755)
ROAi,t	-0.005^**	-0.002
ROAi,t	(-2.375)	(-1.094)
Switchi,t×LEVi,t		0.004^*
Switchi,t×LEVi,t		(1.807)
Switchi,t×ROCi,t		-0.002^**
Switchi,t×ROCi,t		(-2.427)
Switchi,t×CFi,t		0.048^***
Switchi,t×CFi,t		(4.651)
Switchi,t×DIRSi,t		-0.030^***
Switchi,t×DIRSi,t		(-2.791)
Switchi,t×LARSi,t		-0.076^***
Switchi,t×LARSi,t		(-4.463)
Switchi,t×BIG4i,t		0.013
Switchi,t×BIG4i,t		(1.021)
Switchi,t×SIZEi,t		-0.010^***
Switchi,t×SIZEi,t		(-4.153)
Switchi,t×ROAi,t		-0.019^***
Switchi,t×ROAi,t		(-3.564)
Constant	0.211^***	0.166^***
Constant	(10.112)	(7.515)
Industry	Yes	Yes
Year	Yes	Yes
N	30341	30341
Pseudo R2	0.3617	0.3582
Note: *, , and * symbolize two-tailed statistical significance at 1%, 5%, and 10% levels, correspondingly.

From Eq. (1), we derive the values for the opinion shopping variable (OpnShop). The data shows that the mean and median values of OpnShopi,t are 0.008 and 0.005, respectively. Notably, as OpnShop values decrease (i.e., become more negative), there is a higher likelihood of companies improving their audit opinions, particularly when they switch auditors in the presence of factors such as financial leverage (Switchi,t×LEVi,t = 0.004*) and return on capital (Switchi,t×ROCi,t = -0.002**). Conversely, when OpnShop values increase (i.e., become more positive), this likelihood diminishes, indicating that firms may be less inclined to switch auditors to improve audit outcomes under these conditions. Additionally, our analysis reveals a significant negative coefficient for the interaction between Switch and past modified audit opinions (Switchi,t×MAOLagi,t = -0.100***), suggesting that companies may strategically leverage auditor switching to secure more favorable audit opinions, particularly when they have a history of MAOs.

4.3. Results for H1

Table 5 presents the probit regression results for the "Analyst Coverage and Opinion Shopping Model" as outlined in Eq. (3). This model explores the relationship between analyst coverage and auditor switching decisions (Switchi,t), incorporating both continuous and dummy variables for analyst coverage.

In the first column, which excludes the interaction term between Analyst Coveragei,t and OpnShopi,t, the negative coefficient on OpnShopi,t (− 4.169***) suggests a significant propensity for companies to switch auditors, particularly when the likelihood of receiving unclean audit opinions is high. This finding underscores the strategic manipulation of auditor relationships by companies to secure more favorable audit outcomes.

When the interaction term OpnShopi,t × Analysti,t is included in the second column, the model reveals an even stronger negative coefficient (− 2.564**). This pronounced negative coefficient indicates that analyst coverage intensifies the tendency of companies to switch auditors in pursuit of favorable audit opinions. The significant interaction effect suggests that while analyst coverage is intended to enhance market transparency, it can paradoxically motivate firms to change auditors as a strategy to influence audit results.

Furthermore, the inclusion of 'Analyst Presence' as a dummy variable provides additional insight into how analyst coverage affects auditor-switching decisions related to opinion shopping. The interaction term OpnShopi,t × Analyst Presencei,t also shows a significant negative coefficient (− 2.492**), indicating that companies under analyst scrutiny are more likely to switch auditors to obtain favorable audit opinions than those without such scrutiny.

Table 5

Analyst tracking and opinion shopping model
Variables	(1)	(2)	(3)
Variables	Switch_i,t	Switch_i,t	Switch_i,t
OpnShopi,t	-4.169***	-6.406***	-4.970***
OpnShopi,t	(-6.849)	(-7.521)	(-6.572)
Analysti,t		-0.084***
Analysti,t		(-4.814)
OpnShopi,t×Analysti,t		-2.564***
OpnShopi,t×Analysti,t		(-3.897)
Analyst Presencei,t			-0.093**
Analyst Presencei,t			(-2.422)
OpnShopi,t×Analyst Presencei,t			-2.492**
OpnShopi,t×Analyst Presencei,t			(-2.124)
LEVi,t	0.506***	0.479***	0.494***
LEVi,t	(5.687)	(5.320)	(5.481)
ROCi,t	-0.101**	-0.085**	-0.096**
ROCi,t	(-2.395)	(-1.997)	(-2.271)
CFi,t	-0.847***	-0.782***	-0.829***
CFi,t	(-3.573)	(-3.252)	(-3.474)
DIRSi,t	-0.915***	-0.886***	-0.901***
DIRSi,t	(-5.939)	(-5.614)	(-5.709)
LARSi,t	-0.067	-0.141	-0.092
LARSi,t	(-0.603)	(-1.243)	(-0.811)
BIG4i,t	0.168**	0.184**	0.168**
BIG4i,t	(2.296)	(2.497)	(2.277)
SIZEi,t	-0.055***	-0.011	-0.039**
SIZEi,t	(-3.557)	(-0.638)	(-2.321)
ROAi,t	-0.044	-0.038	-0.042
ROAi,t	(-1.533)	(-1.328)	(-1.459)
Constant	-0.747**	-1.522***	-1.014***
Constant	(-2.191)	(-3.938)	(-2.800)
Industry	Yes	Yes	Yes
Year	Yes	Yes	Yes
N	37,774	37,774	37,774
Wald chi2	1280.312	1331.502	1298.415
Pseudo R2	0.043	0.045	0.044
Note: *, , and * symbolize two-tailed statistical significance at 1%, 5%, and 10% levels, correspondingly.

4.4. Results for H2a and H2b

Table 6 presents the analysis for Hypothesis 2, which investigates how political connections influence the relationship between analyst coverage and audit opinion shopping. In Column 1, the interaction term OpnShopi,t × Analysti,t × SOERi,t shows a significant negative coefficient (-0.032***). This suggests that political connections, as measured by SOERi,t, intensify the relationship between analyst coverage and the propensity for companies to engage in audit opinion shopping. The results indicate that firms with political connections are more likely to leverage their influence to obtain favorable audit opinions when under analyst scrutiny. Column 2 further reinforces this finding, with the interaction term OpnShopi,t × Analyst Presencei,t × SOERi,t also showing a significant negative coefficient (-0.119***). This result implies that the mere presence of analysts, combined with political connections, amplifies companies' tendency toward audit opinion shopping. In Column 3, the interaction term OpnShopi,t × Analysti,t × PCi,t has a negative and significant coefficient (-0.027*), indicating that political connections at the executive level (PCi,t) similarly strengthen the relationship between analyst coverage and audit opinion shopping. Column 4 supports this conclusion with the interaction term OpnShopi,t × Analyst Presencei,t × PCi,t which also shows a significant negative coefficient (-0.384**). Overall, the results across all columns consistently validate Hypothesis 2a. Both forms of political connections, as measured by SOERi,t and PCi,t, enhance the relationship between analyst coverage whether measured by presence or intensity and the increased likelihood of companies engaging in audit opinion shopping.

Table 6

Factors affecting partner-level opinion shopping
Variables	(1)	(2)	(3)	(4)
Variables	Switch_i,t	Switch_i,t	Switch_i,t	Switch_i,t
OpnShopi,t	-6.263***	-4.752***	-2.002**	-2.112**
OpnShopi,t	(-7.284)	(-6.263)	(-2.078)	(-2.141)
Analysti,t	-0.074***		-0.098^***
Analysti,t	(-4.172)		(-4.862)
OpnShopi,t×Analysti,t	-2.458***		-0.423*
OpnShopi,t×Analysti,t	(-3.671)		(-1.795)
OpnShopi,t×SOERi,t	0.046	0.038*
OpnShopi,t×SOERi,t	(1.470)	(1.710)
OpnShopi,t×Analysti,t × SOERi,t	-0.032***
OpnShopi,t×Analysti,t × SOERi,t	(-3.496)
Analyst Presencei,t		-0.084**		-0.074*
Analyst Presencei,t		(-2.180)		(-1.706)
OpnShopi,t×Analyst Presencei,t		-2.358**		-1.479
OpnShopi,t×Analyst Presencei,t		(-2.009)		(-1.155)
OpnShopi,t×Analyst Presencei,t×SOERi,t		-0.119***
OpnShopi,t×Analyst Presencei,t×SOERi,t		(-2.851)
OpnShopi,t×Analysti,t×PCi,t			-0.027*
OpnShopi,t×Analysti,t×PCi,t			(-1.728)
OpnShopi,t×PC i,t			0.560	0.480
OpnShopi,t×PC i,t			(0.368)	(0.302)
OpnShopi,t×Analyst Presencei,t×PCi,t				-0.384**
OpnShopi,t×Analyst Presencei,t×PCi,t				(-2.292)
PC i,t			-0.206^***	-0.223^***
PC i,t			(-4.727)	(-5.278)
SOERi,t	0.005***	0.005***	1.132^***	1.112^***
SOERi,t	(6.079)	(6.476)	(5.773)	(5.705)
LEVi,t	0.486***	0.491***	-0.013	-0.014
LEVi,t	(5.373)	(5.421)	(-1.037)	(-1.056)
ROCi,t	-0.084**	-0.092**	0.001	0.002
ROCi,t	(-1.974)	(-2.190)	(0.327)	(0.411)
CFi,t	-0.786***	-0.818***	0.042	0.049
CFi,t	(-3.250)	(-3.407)	(0.426)	(0.512)
DIRSi,t	-0.774***	-0.764***	0.047	0.016
DIRSi,t	(-4.878)	(-4.814)	(0.525)	(0.162)
LARSi,t	-0.340***	-0.298**	0.299^*	0.488^***
LARSi,t	(-2.833)	(-2.495)	(1.652)	(2.769)
BIG4i,t	0.206***	0.195***	0.145^*	0.119
BIG4i,t	(2.796)	(2.642)	(1.676)	(1.378)
SIZEi,t	-0.040**	-0.065***	-0.022	0.007
SIZEi,t	(-2.150)	(-3.781)	(-0.866)	(0.275)
ROAi,t	-0.047	-0.051*	-0.035	-0.027
ROAi,t	(-1.583)	(-1.708)	(-1.012)	(-1.608)
Constant	-1.013**	-0.552	-1.151^*	-1.573^**
Constant	(-2.562)	(-1.498)	(-1.786)	(-2.469)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
N	37,774	37,774	37,774	37,774
Wald chi2	1388.29	1371.628	1174.099	1152.692
Pseudo R2	0.046	0.045	0.048	0.048
Note: *, , and * symbolize two-tailed statistical significance at 1%, 5%, and 10% levels, correspondingly.

5.1. Impact of Analyst Announcements on Audit Opinion Shopping

Drawing from Griffin et al. (2003) and Li and Ramesh (2009), our study investigates the effect of analyst announcements on audit opinion shopping within China's A-share market. We employ event study methodology to calculate excess returns, defined as the difference between a firm's buy-and-hold stock return and the buy-and-hold return of the CSI 300 index during specific event windows around the analyst announcement (days 0 to 2, -3 to 3, and − 5 to 5).

To ensure relevance and impact, we focus exclusively on reports issued by the most influential analysts in each industry, as identified by the "New Fortune Best Analyst1" rankings. This approach is necessary due to the large number of analysts and reports, making it impractical to determine consistent event windows otherwise. By selecting reports from top-ranked analysts, we capture significant market reactions due to their high regard and substantial influence on investors.

Table 7 presents our findings, showing that politically connected firms exhibit significant increases in excess returns during these windows, particularly immediately following analyst announcements. The cumulative abnormal returns (CAR) for firms with strong political connections are 0.932 for the [0, 2] window, 0.256 for the [-3, + 3] window, and 0.859 for the [-5, + 5] window. These results suggest that such firms strategically leverage positive analyst coverage to enhance their market image, which may lead to increased audit opinion shopping to reflect this positive market sentiment.

In contrast, firms with weaker political connections show milder market reactions within the same windows, with CARs of -0.295, 0.413, and − 0.292, respectively. This indicates a lesser impact of analyst announcements on their market perception, potentially due to reduced reliance on political influence.

For the overall sample, the results show CARs of 0.134 for the [0, 2] window, 0.834 for the [-3, + 3] window, and 0.576 for the [-5, + 5] window. This suggests that, on average, firms experience positive excess returns around analyst announcements, although the effect is less pronounced than for politically connected firms.

These findings underscore the significant role political connections play in shaping company responses to analyst reports in the context of audit opinion shopping. Firms with strong political ties seem more motivated to use favorable analyst coverage to bolster their market image, potentially increasing their propensity for audit opinion shopping. This aligns with the notion that politically connected firms, under heightened scrutiny, actively use their political capital to influence market perceptions effectively (Liu et al., 2021).

Table 7

Market reactions to analyst announcements in high and low political connection groups
Event period	All (Simulated Obs. = 852)		High Political Connection (Obs. = 644)		Low Political Connection (Obs. = 208)
	CAR	t-value	CAR	t-value	CAR	t-value
[0, 2]	0.134***	(8.921)	0.932***	(18.480)	-0.295	(-1.360)
[-3, + 3]	0.834***	(7.807)	0.256***	(15.526)	0.413*	(1.691)
[-5, + 5]	0.576***	(5.504)	0.859***	(9.701)	-0.292	(-1.262)
Note: 1) This table presents the descriptive statistics of excess returns around the announcement dates of reports by star analysts in China's A-share market. Excess returns are the difference between a firm's buy-and-hold stock return and the return of the CSI 300 index over the event windows of days 0 to 2, -3 to 3, and − 5 to 5, with day 0 being the analyst report publication date. 2) Panel A shows results for firms with high political connections, while Panel B focuses on firms with low political connections. This differentiation provides insights into how political connections might influence market responses to analyst announcements and audit opinion shopping behaviors. 3) Significance levels are denoted by * (1%), (5%), and * (10%) for the t-tests.

In Table 8, we present a multiple regression analysis examining the excess returns around analyst announcement dates. We utilize quarterly data from firms with analyst coverage and a control group without such coverage. The dependent variable in our regression is the excess return around the analyst report publication date, measured over three different event windows.

The variable Announcement is a dummy variable assigned a value of 1 for firm-quarters coinciding with the analyst report publication and 0 otherwise. PolConnecti,t indicates whether a firm has high political connections2. We include an interaction term Announcement×PConnecti,t to capture any additional excess returns attributable to political connections at the time of the analyst announcements.

Consistent with prior research, we control for factors influencing a firm’s excess return around these events. These factors include market capitalization (MarketCapi,t), book-to-market value (BTMi,t), institutional ownership (InstOwni,t), and the 252-day standard deviation of the firm’s return before the announcement date (ReturnSDi,t). All financial variables are measured at the end of the fiscal quarter immediately preceding the announcement date. Definitions of these variables are detailed in the Appendix.

The regression results demonstrate a significant positive impact of analyst announcements on market reactions, particularly for companies with political connections. This suggests that political connections may amplify the effect of analyst announcements on stock prices and returns.

Table 8

Impact of analyst announcements and political connections on market reactions
Variables	[0, 2]	[-3, + 3]	[-5, + 5]
Announcement	0.013*	0.033***	0.014*
Announcement	(1.953)	(2.871)	(2.020)
PolConnecti,t	0.151*	0.104**	0.103
PolConnecti,t	(1.701)	(2.420)	(1.580)
Announcement×PolConnecti,t	0.014***	0.024*	0.009***
Announcement×PolConnecti,t	(2.020)	(1.806)	(3.274)
MarketCapi,t	1.145***	1.145***	1.145***
MarketCapi,t	(550.029)	(550.031)	(550.022)
BTMi,t	-0.174***	-0.174***	-0.174***
BTMi,t	(-25.670)	(-25.703)	(-25.695)
InstOwni,t	0.104***	0.104***	0.104***
InstOwni,t	(27.054)	(27.037)	(27.042)
ReturnSDi,t	-0.029***	-0.029***	-0.029***
ReturnSDi,t	(-7.877)	(-7.876)	(-7.879)
Constant	-1.056***	-1.057***	-1.057***
Constant	(-22.772)	(-22.813)	(-22.804)
N	1,112,923	1,112,923	1,112,923
Firm FE	Yes	Yes	Yes
Note:

1) This table presents the regression results analyzing the market response to analyst announcements and their impact on audit opinion shopping. The regression model assesses excess returns over three distinct event windows: days 0 to 2, days − 3 to 3, and days − 5 to 5. These windows were selected to capture the immediate and extended market reactions around the announcement dates, with day 0 marking the release of the analyst report.

2) The dependent variables for the regression models are the excess returns observed during these event windows. Excess returns are defined as a firm’s buy-and-hold stock return minus the buy-and-hold return of a representative Chinese stock market index, such as the CSI 300, over the respective event window.

3) The t-values, provided in parentheses, indicate the statistical significance of the coefficients. Standard errors are adjusted for heteroskedasticity and firm clustering. Significance levels are denoted by ***, **, and *, corresponding to the 1%, 5%, and 10% levels. The model also includes firm fixed effects to control for unobserved heterogeneity.

4) The sample includes observations from China's A-share market starting in 2004, when analyst tracking data became increasingly available and robust. The analysis aims to show how analyst reports, especially from prominent analysts, affect short-term market reactions and potentially influence corporate behaviors regarding audit opinion shopping.

5.2 Effect of Political Leadership Turnover on Audit Opinion Shopping

Prior research suggests that local officials, driven by political promotion tournaments, often use their influence to implement financial policies that favor strategic industries, particularly by manipulating credit markets to direct bank loans (Bao et al., 2016; Cao et al., 2019; Li and Zhou, 2005). However, there is growing concern that the uncertainty caused by officials' turnover may negatively impact economic outcomes, increasing local debt risks, nonperforming loans, and corporate pollution, while reducing investment (Chen et al., 2005; Zhong et al., 2019). Given these dynamics, it becomes crucial to understand how the background of local officials whether they are promoted from within the region or parachuted in from another affects audit opinion shopping, a key area where political influence may be exerted.

To analyze this, we create three dummy variables named LocalPromote17, LocalPromote18, and LocalPromote19, corresponding to the 17th, 18th, and 19th National Congresses of the Communist Party of China (CPC)3, respectively. Each variable indicates whether the provincial party secretary during that period was a locally promoted official (coded as 1) or parachuted in from another region (coded as 0). Our analysis is based on the location of the listed companies, examining whether the nature of the provincial leadership influences the propensity for audit opinion shopping in different political periods.

Table 9 examines how political leadership turnover influences audit opinion shopping, focusing on the interaction between local political promotion, analyst coverage, and audit behavior. The results show that for the 17th, 18th, and 19th National Congress periods, the interaction terms LocalPromote × Analysti,t × OpnShopi,t are all significantly negative. This indicates that firms with locally promoted officials are more likely to engage in audit opinion shopping when they are under analyst coverage, suggesting that these firms may use their political connections and the pressure from analysts to secure favorable audit outcomes. Similarly, the interaction terms involving Analyst Presence show negative coefficients, particularly during the 18th and 19th National Congress periods. This further suggests that the mere presence of analysts, combined with local political leadership, enhances the likelihood of audit opinion shopping.

Table 9

Analysis of the Impact of Political Leadership Turnover on Audit Opinion Shopping
Variables	(1)	(2)
Variables	Switch_i,t	Switch_i,t
OpnShopi,t	-9.859^***	-9.670^***
OpnShopi,t	(-11.907)	(-10.970)
LocalPromote17×Analysti,t×OpnShopi,t	-0.025^**
LocalPromote17×Analysti,t×OpnShopi,t	(-2.015)
LocalPromote18×Analysti,t×OpnShopi,t	-0.823^***
LocalPromote18×Analysti,t×OpnShopi,t	(-2.520)
LocalPromote19×Analysti,t×OpnShopi,t	-0.119^**
LocalPromote19×Analysti,t×OpnShopi,t	(-2.064)
LocalPromote17×OpnShopi,t	1.679	2.067
LocalPromote17×OpnShopi,t	(0.864)	(0.974)
LocalPromote18×OpnShopi,t	4.846^*	7.193^**
LocalPromote18×OpnShopi,t	(1.929)	(2.562)
LocalPromote19×OpnShopi,t	-1.093	-0.934
LocalPromote19×OpnShopi,t	(-0.395)	(-0.291)
Analysti,t × OpnShopi,t	0.460
Analysti,t × OpnShopi,t	(0.868)
LocalPromote17×Analyst Presencei,t × OpnShopi,t		-1.529^*
LocalPromote17×Analyst Presencei,t × OpnShopi,t		(-1.441)
LocalPromote18×Analyst Presencei,t × OpnShopi,t		-4.909^*
LocalPromote18×Analyst Presencei,t × OpnShopi,t		(-1.789)
LocalPromote19×Analyst Presencei,t × OpnShopi,t		-0.538^**
LocalPromote19×Analyst Presencei,t × OpnShopi,t		(-2.305)
Analysti,t	-0.100^***
Analysti,t	(-4.941)
LocalPromote17	1.066^***	1.026^***
LocalPromote17	(10.757)	(10.349)
LocalPromote18	0.827^***	0.780^***
LocalPromote18	(9.154)	(8.669)
LocalPromote19	-0.099	-0.123
LocalPromote19	(-0.989)	(-1.223)
Analyst Presencei,t×OpnShopi,t		-0.833
Analyst Presencei,t×OpnShopi,t		(-0.679)
Analyst Presencei,t		-0.078
Analyst Presencei,t		(-1.591)
LEVi,t	0.030^**	0.028^**
LEVi,t	(2.375)	(2.306)
ROCi,t	-0.021^***	-0.023^***
ROCi,t	(-8.019)	(-9.552)
CFi,t	0.455^***	0.445^***
CFi,t	(5.031)	(3.808)
DIRSi,t	-0.252^***	-0.328^***
DIRSi,t	(-2.799)	(-3.769)
LARSi,t	-0.260^**	-0.340^***
LARSi,t	(-2.112)	(-2.790)
BIG4i,t	0.285^***	0.270^***
BIG4i,t	(3.417)	(3.234)
SIZEi,t	-0.073^***	-0.111^***
SIZEi,t	(-3.810)	(-6.137)
ROAi,t	-0.182^***	-0.204^***
ROAi,t	(-5.064)	(-0.122)
Constant	-0.293	0.555
Constant	(-0.632)	(1.274)
Industry	Yes	Yes
Year	Yes	Yes
N	30341	30341
Wald chi2	1146.532	1120.319
Pseudo R2	0.048	0.047
Note: *, , and * symbolize two-tailed statistical significance at 1%, 5%, and 10% levels, correspondingly.

[1] The "New Fortune Best Analyst" ranking, known as the "Oscars of the brokerage industry," is China's most influential analyst ranking, launched in 2003. It annually recognizes the top analysts across various fields, based on votes from institutional investors such as public funds, insurance companies, and private equity firms.

[2] PolConnecti,t indicates whether a firm has high political connections. We classify a firm as having high political connections if its SOERi,t value is above the median; otherwise, it is considered to have low political connections.

[3] Our sample covers the period from 2004 to 2022, which is why we focus on the 17th, 18th, and 19th National Congresses during this time frame.

6.1. Addressing endogeneity in analyst tracking and audit opinion shopping

Addressing endogeneity, our study tackles the bidirectional causality between analyst tracking and audit opinion shopping, along with the intertwined effects of political connections. We employ a two-stage least squares (2SLS) regression, using inclusion in the Shanghai and Shenzhen 300 Index (H&S300) as an instrumental variable, following Shih et al. (2021) 's methodology. This approach mitigates biases by leveraging H&S300's criteria, which are unrelated to audit behaviors but linked to increased analyst attention. Our analysis confirms that H&S300's market prominence, which correlates with analyst tracking yet is exogenous to audit opinion shopping, serves as a robust instrument.

Our 2SLS regression, based on Larcker and Rusticus (2010), further examines the interplay between analyst attention, political connections, and audit opinion shopping. The instrumental variable's validity is supported by strong F-statistics and an exogenous Hansen J statistic p-value, as shown in Table 10. The results confirm that firms under greater analyst scrutiny, especially those with strong political ties, are more inclined to seek favorable audit outcomes.

Table 10

2SLS results
Variables	First stage	Second stage	First stage	Second stage
Variables	DV = H&S300	DV = Switch	DV = H&S300	DV = Switch
OpnShopi,t	-5.929***	-0.396	-5.801***	-4.170**
OpnShopi,t	(-20.021)	(-0.574)	(-19.491)	(-2.503)
Analysti,t		-0.015		-0.010
Analysti,t		(-1.059)		(-0.708)
SOERi,t			0.025**	0.003***
SOERi,t			(3.253)	(5.767)
OpnShopi,t×SOERi,t			0.057	0.062*
OpnShopi,t×SOERi,t			(1.630)	(1.810)
OpnShopi,t×Analysti,t	-4.037***	-2.675***	-4.119***	-2.477***
OpnShopi,t×Analysti,t	(-14.510)	(-5.874)	(-14.656)	(-4.731)
OpnShopi,t×Analysti,t×SOERi,t			-0.0425***	-0.028***
OpnShopi,t×Analysti,t×SOERi,t			(-7.341)	(-4.605)
LEVi,t	0.001***	-0.000	0.001***	-0.000
LEVi,t	(5.010)	(-1.421)	(5.057)	(1.228)
ROCi,t	-0.003	-0.003*	-0.000	-0.000*
ROCi,t	(-0.976)	(-2.111)	(-0.928)	(-2.117)
CFi,t	0.007	0.009***	0.007	0.009***
CFi,t	(1.901)	(4.272)	(1.854)	(4.067)
DIRSi,t	0.012	-0.001	0.011	0.001
DIRSi,t	(1.692)	(-0.221)	(1.542)	(0.179)
LARSi,t	-0.047***	0.039***	-0.042***	0.014
LARSi,t	(-5.871)	(3.302)	(-5.151)	(1.297)
BIG4i,t	0.121***	0.0133	0.118***	0.0170*
BIG4i,t	(13.794)	(1.580)	(13.494)	(2.040)
SIZEi,t	0.115***	0.006	0.115***	0.002
SIZEi,t	(71.447)	(1.108)	(68.810)	(0.294)
ROAi,t	-0.043	-0.016	-0.038	-0.030
ROAi,t	(-1.650)	(-0.474)	(-1.461)	(-0.884)
Constant	-2.130***	-0.021	-2.131***	0.058
Constant	(-57.728)	(-0.219)	(-56.308)	(0.544)
Industry	Yes	Yes	Yes	Yes
Year	Yes	Yes	Yes	Yes
N	37,774	37,774	37,774	37,774
R-squared	0.321	0.244	0.342	0.166
Cragg–Donald Wald F statistic (weak identification test)	1066.333		1056.206
Kleibergen–Paap rk Wald F statistic (weak identification test)	565.340		559.692
Hensen J statistic p-value (overidentification test of all instruments)	0.185		0.195
Note: 1) This table presents the results of IV regressions that address the endogeneity of the variables of interest in our study: analyst tracking, audit opinion shopping, and political connections. The instrumental variable used is the inclusion in the Shanghai and Shenzhen 300 Index (H&S300) for the number of analysts tracking a firm. The first stage regression (Model 1) shows the relationship between the instrumental variable and the endogenous variable of analyst tracking. Model 2 provides results for the second stage regressions (2SLS) where the dependent variable is Switch. Both models include industry and year controls. The R-squared values indicate the explanatory power of each model. The table also includes the Cragg–Donald Wald F statistic and the Kleibergen–Paap rk Wald F statistic for weak identification tests, as well as the Hansen J statistic p-value for the overidentification test of all instruments. Standard errors are clustered at the firm and year level to account for serial correlation. Significance levels are denoted by , , and ** for the 10%, 5%, and 1% levels, respectively. Definitions for the variables are provided in the Appendix.

This study examines the relationship between analyst coverage and audit opinion shopping within the context of Chinese-listed firms, specifically focusing on how political connections and political leadership turnover influence this dynamic. Our findings reveal several critical insights.

First, we find that analyst coverage significantly increases the likelihood of firms engaging in audit opinion shopping. This behavior is particularly pronounced in firms with strong political connections, where the pressure from analyst scrutiny appears to motivate management to seek more favorable audit outcomes. These results add to the existing literature by highlighting the nuanced role that analyst monitoring can play in corporate governance where, instead of mitigating opportunistic behavior, it can inadvertently encourage it under certain conditions.

Second, through a detailed event study, we analyze the impact of analyst announcements on audit opinion shopping within China’s A-share market. By focusing on reports from top-ranked analysts, our study captures significant market reactions and demonstrates that influential analyst coverage can heighten the pressure on firms to secure favorable audit outcomes. This finding contributes to the literature by illustrating how market responses to analyst reports can influence audit practices, particularly during periods of heightened financial scrutiny.

Third, our examination of political leadership turnover during the 17th, 18th, and 19th National Congresses reveals that changes in local political leadership, especially the promotion of local officials, further exacerbate firms’ tendencies to engage in audit opinion shopping. This aspect of the study provides a deeper understanding of how political dynamics at the regional level interact with corporate audit practices, offering new insights into the role of political stability and leadership continuity in shaping corporate behavior.

These findings carry significant policy and practical implications. Regulators and market participants should be vigilant in monitoring firms that are subject to both intense analyst scrutiny and strong political connections, as these firms are more prone to engage in audit opinion shopping. Ensuring the independence and integrity of the audit process in such environments is critical for maintaining the reliability of financial reporting. Moreover, policymakers should consider the dual impact of market and political pressures when designing regulatory frameworks, to better protect against manipulative practices that undermine investor confidence and market stability.

Finally, this study has some limitations, such as the focus on Chinese-listed firms, which may limit the generalizability of the findings to other markets. Future research could explore similar dynamics in different institutional contexts or consider additional factors, such as cultural influences or variations in audit quality, to provide a more comprehensive understanding of the factors that drive audit opinion shopping.

Author Contribution

Xinwei Wu: Writing – review & editing, Writing – original draft, Methodology, Formal analysis, Data. Han Sun: Writing – review & editing, Supervision, Investigation. Hyoung-Goo Kang: Validation, Supervision, Methodology, Formal analysis. JaeHo Lee: Writing – review & editing, Validation, Supervision, Resources.

Adhikari A, Derashid C, Zhang H (2006) Public policy, political connections, and effective tax rates: Longitudinal evidence from Malaysia. J Account Public Policy 25(5):574–595. https://doi.org/10.1016/j.jaccpubpol.2006.07.001
Asquith P, Mikhail MB, Au AS (2005) Information content of equity analyst reports. J Financ Econ 75(2):245–282. https://doi.org/10.1016/j.jfineco.2004.01.002
Ayres DR, Campbell JL, Chyz JA, Shipman JE (2019) Do financial analysts compel firms to make accounting decisions? Evidence from goodwill impairments. Rev Acc Stud 24(4):1214–1251. https://doi.org/10.1007/s11142-019-09512-0
Bao X, Johan S, Kutsuna K (2016) Do political connections matter in accessing capital markets? Evidence from China. Emerg Markets Rev 29:24–41. https://doi.org/10.1016/j.ememar.2016.08.009
Bliss MA, Gul FA, Majid A (2011) Do political connections affect the role of independent audit committees and CEO duality? Some evidence from Malaysian audit pricing. J Contemp Acc Econ 7(2):82–98. https://doi.org/10.1016/j.jcae.2011.10.001
Bonner SE, Hugon A, Walther BR (2007) Investor Reaction to Celebrity Analysts: The Case of Earnings Forecast Revisions. J Accounting Res 45(3):481–513. https://doi.org/10.1111/j.1475-679x.2007.00245.x
Cao C, Dong Y, Hou W, Liu Y, Qian X (2019) Provincial official turnover and bank loans. Pac-Basin Financ J 57:101202. https://doi.org/10.1016/j.pacfin.2019.101202
Chen Y, Li H, Zhou L-A (2005) Relative performance evaluation and the turnover of provincial leaders in China. Econ Lett 88(3):421–425. https://doi.org/10.1016/j.econlet.2005.05.003
Cho M, Ki E, Kwon SY (2017) The effects of accruals quality on audit hours and audit fees. J Acc Auditing Finance 32(3):372–400. https://doi.org/10.1177/0148558X15625622
Chu KCK, Zhai WHS (2021) Distress risk puzzle and analyst forecast optimism. Rev Quant Finan Acc 57:429–460. https://doi.org/10.1007/s11156-020-00950-5
Czernkowski R, Green W, Wang Y (2010) The value of audit qualifications in China. Managerial Auditing J 25(5):404–426. https://doi.org/10.1108/02686901011041866
Davidson III, Jiraporn WN, P., DaDalt P (2006) Causes and consequences of audit shopping: an analysis of auditor opinions, earnings management, and auditor changes. Q J Bus Econ 69–87. https://doi.org/10.1108/02686900610680589
Dopuch N, Holthausen RW, Leftwich RW (1986) Abnormal stock returns associated with media disclosures of ‘subject to’ qualified audit opinions. J Account Econ 8(2):93–117. https://doi.org/10.1016/0165-4101(86)90008-3
Fan JP, Wong TJ, Zhang T (2007) Politically connected CEOs, corporate governance, and Post-IPO performance of China's newly partially privatized firms. J Financ Econ 84(2):330–357. https://doi.org/10.1016/j.jfineco.2006.03.008
Ghosh A, Moon D (2005) Auditor Tenure and Perceptions of Audit Quality. Acc Rev 80(2):585–612. https://doi.org/10.2308/accr.2005.80.2.585
Green W, Czernkowski R, Wang Y (2009) Special treatment regulation in China: potential unintended consequences. Asian Rev Acc 17(3):198–211. https://doi.org/10.1108/13217340911002667
Griffin JM, Harris JH, Topaloglu S (2003) The dynamics of institutional and individual trading. J Finance 58(6):2285–2320. https://doi.org/10.1046/j.1540-6261.2003.00608.x
Gul FA, Qiu A, Kim JB (2007) Political economy, client economic importance and auditor independence: some evidence from Chinese non-big 4 auditors. Client Economic Importance and Auditor Independence: Some Evidence from Chinese Non-Big 4 Auditors, https://doi.org/10.2139/ssrn.971233
He K, Pan X, Tian GG (2017) Political connections, audit opinions, and auditor choice: Evidence from the ouster of government officers. Auditing: J Pract Theory 36(3):91–114. https://doi.org/10.2308/ajpt-51564
He X, Pittman JA, Rui OM, Wu D (2017) Do social ties between external auditors and audit committee members affect audit quality? Acc Rev 92(5):61–87. https://doi.org/10.2308/accr-51598
Kallias A, Kallias K, Zhang S (2022) Can we trust the accounting discretion of firms with political money contributions? Evidence from U.S. IPOs. J Account Public Policy 41(6):106999. https://doi.org/10.1016/j.jaccpubpol.2022.106999
Khan A, Mihret DG, Muttakin MB (2016) Corporate political connections, agency costs and audit quality. Int J Acc Inform Manage 24(4):357–374. https://doi.org/10.1108/IJAIM-01-2016-0003
Kliger D, Kudryavtsev A (2013) Volatility expectations and the reaction to analyst recommendations. J Econ Psychol 37(37):1–6. https://doi.org/10.1016/j.joep.2013.04.003
Kothari SP, Li X, Short JE (2009) The Effect of Disclosures by Management, Analysts, and Business Press on Cost of Capital, Return Volatility, and Analyst Forecasts: A Study Using Content Analysis. Acc Rev 84(5):1639–1670. https://doi.org/10.2308/accr.2009.84.5.1639
Lang MH, Lins KV, Miller DP (2004) Concentrated Control, Analyst Following, and Valuation: Do Analysts Matter Most When Investors Are Protected Least? J Accounting Res 42(3):589–623. https://doi.org/10.1111/j.1475-679x.2004.t01-1-00142.x
Larcker DF, Rusticus TO (2010) On the use of instrumental variables in accounting research. J Account Econ 49(3):186–205. https://doi.org/10.1016/j.jacceco.2009.11.004
Lennox C (2000) Do companies successfully engage in opinion-shopping? Evidence from the UK. J Account Econ 29(3):321–337. https://doi.org/10.1016/S0165-4101(00)00026-7
Lennox C (2002) Opinion shopping and audit committees. J Pract Theory 19:47–66. https://doi.org/10.2308/accr.2001.76.2.239
Li EX, Ramesh K (2009) Market reaction surrounding the filing of periodic SEC reports. Acc Rev 84(4):1171–1208. https://doi.org/10.2308/accr.2009.84.4.1171
Li H, Zhou L-A (2005) Political turnover and economic performance: the incentive role of personnel control in China. J Public Econ 89(9–10):1743–1762. https://doi.org/10.1016/j.jpubeco.2004.06.009
Lin Z, Wang L (2022) Analyst following, financial constraint, and audit opinion shopping: From the perspective of earning management. J Int Financial Manage Acc 34(4):71–96. https://doi.org/10.1111/jifm.12163
Liu Y, Dai W, Liao M, Wei J (2021) Social status and corporate social responsibility: Evidence from Chinese privately owned firms. J Bus Ethics 169:651–672. https://doi.org/10.1007/s10551-019-04350-6
Lu T (2006) Does opinion shopping impair auditor independence and audit quality? J Accounting Res 44(3):561–583. https://doi.org/10.1111/j.1475-679X.2006.00208.x
Lui D, Markov S, Tamayo A (2007) What makes a stock risky? Evidence from sell-side analysts' risk ratings. J Accounting Res 45(3):629–665. https://doi.org/10.1111/j.1475-679X.2007.00242.x
Lui D, Markov S, Tamayo A (2012) Equity analysts and the market's assessment of risk. J Accounting Res 50(5):1287–1317. https://doi.org/10.1111/j.1475-679X.2012.00462.x
Newton NJ (2019) When Analysts Speak, Do Auditors Listen? AUDITING: A Journal of. Pract Theory 38(1):221–245. https://doi.org/10.2308/ajpt-52059
Newton NJ, Persellin JS, Wang D, Wilkins MS (2016) Internal control opinion shopping and audit market competition. Acc Rev 91(2):603–623. https://doi.org/10.2308/accr-51104
Rajan R, Servaes H (1997) Analyst Following of Initial Public Offerings. J Finance 52(2):507–529. https://doi.org/10.1111/j.1540-6261.1997.tb04811.x
Romain Boissin, Sentis P (2012) Long-run performance of IPOs and the role of financial analysts: some French evidence. Eur J Financ 20(2):125–149. https://doi.org/10.1080/1351847x.2012.689773
Shih YC, Wang Y, Zhong R, Ma YM (2021) Corporate environmental responsibility and default risk: Evidence from China. Pac-Basin Financ J 68:101596. https://doi.org/10.1016/j.pacfin.2021.101596
Tantawy SM, Moussa T (2023) The effect of political connections on firms' auditor choice decisions and audit opinions: evidence from Egypt. Asian Rev Acc. https://doi.org/10.1108/ARA-06-2022-0137
Tee CM, Gul FA, Foo YB, Teh CG (2017) Institutional monitoring, political connections and audit fees: Evidence from Malaysian firms. Int J Auditing 21(2):164–176. https://doi.org/10.1111/ijau.12086
Tessema A, Abou-El-Sood H (2022) Audit rotation, information asymmetry and the role of political connections: international evidence. Meditari Account Res 31(5):1352–1385. https://doi.org/10.1108/medar-03-2021-1226
Wahab EAA, Gist WE, Majid WZNA (2014) Characteristics of non-audit services and financial restatements in Malaysia. J Contemp Acc Econ 10(3):225–247. https://doi.org/10.1016/j.jcae.2014.09.003
Wahab EAA, Haron H, Lok CL, Yahya S (2011) Does corporate governance matter? Evidence from related party transactions in Malaysia. Int Corp Gov 14:131–164. https://doi.org/10.1108/S1041-7060(2011)0000014010
Wahab EAA, Zain MM, Rahman RA (2013) Political connections, fees paid to auditors and auditor independence in Malaysia: evidence from going concern audit opinions. Int J Acc Auditing Perform Evaluation 9(2):153–183. https://doi.org/10.1504/IJAAPE.2013.056487
Wiersema MF, Zhang Y (2011) CEO Dismissal: The role of investment analysts. Strateg Manag J 32(11):1161–1182. https://doi.org/10.1002/smj.932
Yang X, Dong L, Nahm A (2023) Mingling business and politics in China–Does it build value? The relationship between political connection and firm performance. J Strategy Managemen. https://doi.org/10.1108/JSMA-12-2022-0260
Zhang L, Wei M (2019) How analyst recommendations respond to corporate uncertainty caused by investment behavior. China Finance Rev Int 10(3):243–269. https://doi.org/10.1108/cfri-05-2019-0046
Zhong W, Lin Y, Gao D, Yang H (2019) Does politician turnover affect foreign subsidiary performance? Evidence in China. J Int Bus Stud 50(7):1184–1212. https://doi.org/10.1057/s41267-019-00229-5

No competing interests reported.

Appendix1.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Assessing the Impact of Analyst Coverage and Political Connections on Audit Opinion Shopping in China's Corporate Sector

Status:

Version 1

Abstract

1. Introduction

2. Theory and Hypothesis Development

2.1 The Influence of Analyst Coverage on Audit Opinion Shopping

2.2 The Moderating Role of Political Connections in Audit Opinion Shopping

3. Research Methodology

3.1 Sample selection

3.2 Variable description and Analysis models

3.2.1 Audit opinion shopping

3.2.2 Analyst Coverage

3.2.3 Political Connections

3.3 Analysis models

4. Empirical results

4.1 Descriptive statistics

4.2. Results from the audit reporting model

4.3. Results for H1

4.4. Results for H2a and H2b

5. Additional analyses

5.1. Impact of Analyst Announcements on Audit Opinion Shopping

5.2 Effect of Political Leadership Turnover on Audit Opinion Shopping

6. The Endogeneity Tests

6.1. Addressing endogeneity in analyst tracking and audit opinion shopping

7. Conclusions and Implication

Declarations

Author Contribution

References

Additional Declarations

Supplementary Files

Status:

Version 1