Tax avoidance and the timing of annual earnings announcements in China

(1)

Tax avoidance and the timing of annual earnings

announcements in China

Name: Yue Xu

Student number: 11388587 Thesis supervisor: dr R. Felleg Date: 26 June 2017

Word count: 14,263

MSc Accountancy & Control, specialization Accountancy Amsterdam Business School

(2)

Statement of Originality

This document is written by student Yue Xu who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

This study examines to what degree tax avoidance influences the timing of the annual earnings announcement in China. This research question is motivated by the growing significance of timeliness of earnings announcements with regards to news and the increasing importance of tax avoidance.

I first confirm that there is a positive relationship between tax avoidance and the delay in the annual earnings announcement, which means that higher tax avoidance significantly leads to longer reporting lag. Then I confirm that there is a significant association between the audit quality and the relationship of tax avoidance and the reporting lag. Next, I find that audit quality is significantly positively associated with the delay in annual earnings announcements.

(4)

1 Introduction

This study examines how the timing of annual earnings announcement is affected by tax avoidance in China. This research question is motivated by the growing significance of timeliness of earnings announcements with regards to news and the increasing importance of tax avoidance. First of all, accounting standards for disclosure emphasize on the timeliness of accounting information. While corporations want to disclose information at an optimal timing, it is uncertain whether managers make the best strategy for firms in activities like earnings announcements (Chen and Mohan, 1994). There are different indicators for fast as well as delayed release of bad news (Kothari et al., 2008). When facing the exercise price of the options or the litigation risk, managers will be motivated to disclose bad news early (Aboody and Kasznik, 2000). However, there are also incentives for managers to not release the bad news. There is a saying of ‘good news early, bad news late’ at the information market (Baginski et al., 2002). This phenomenon implies that in order to prevent bad news from being worse when there are unexpected negative earnings, managers choose to delay the timing of disclosing earnings information (Trueman, 1990). Such delay in releasing the earnings information gives the managers more benefits versus costs (Begley, 1998). Therefore, it is of interest to study the determinants of the delay in annual earnings disclosure for both the regulators who are interested in timely disclosed information and the investors who anticipate the latest news from the annual disclosure (Kim and Son 2014).

(6)

In parallel to the growing significance of timeliness of earnings announcements, tax avoidance has become a hot topic in both academic accounting research and practical operation (Dyreng et al., 2005; Hanlon and Heitzman, 2010). Studies show that on the basis of agency theory which details the conflicts of interests between managers and the principals, the activities of tax avoidance can further promote managerial opportunism (Desai and Dharmapala, 2006). Because of ‘good news early, bad news late’, managers undertake tax avoidance activities to hide bad news, which contributes to investors’ misunderstanding of the performance of the firm (Desai and Dharmapala, 2006). By using tax planning, managers can manipulate earnings (Kim et al., 2011). Therefore, it is of great interest to study the level of tax avoidance. Since managers determine not only tax avoidance but also the timing of the earnings announcements per year, there must be a relationship between them.

Apart from the above two motivations, the third one is the gradually improved role that China plays in the global market. Many companies are interested in exploring investments and partnerships in China (Campbell and Wang, 2014). Therefore, the way in which Chinese firms manage taxes and disclose information is important. And the Chinese setting enables me to examine the relationship better. The fiscal year end is December 31 in China. The filling deadline is April 30 in China. This helps me to have a clear exploration on the relationship.

Prior studies have tested how managers’ discretion influences the delay in annual earnings announcements in general (Kim and Son, 2014; Chai and Tung 2002). These

(7)

studies document that if there is good news, earnings will be disclosed earlier. Furthermore, whether the disclosure of the annual earnings announcements is delayed depends on the anticipation of bad news in both the Chinese and American markets. Moreover, the earlier researches have also proved the existence of the relationship between tax avoidance and the managers’ behavior (Desai and Dharmapala, 2006). These studies imply that high-powered incentives can influence corporate tax avoidance. My study is different because I emphasis on the consequence of tax avoidance rather than the reasons for tax avoidances. Some other studies have suggested that there is an association between tax avoidance and the timing of annual earnings announcements. And evidence is given by Crabtree and Kubick (2014), showing this relationship in the United States. They have discovered that there is a certain degree of association between tax avoidance and the timing of annual earnings disclosure, but have found no sign of this relationship. I further examine the relationship, and based on the agency theory and the settings in China, I believe that there is a positive relationship between tax avoidance and the delay in the annual earnings announcement in China.

Therefore, I examine how tax avoidance affects timing of the annual earnings announcement. I choose the book-tax differences (BTDs) in China because according to Chinese GAAP and IFRS, the expense recognition is relatively conservative in the tax laws (Tang and Firth, 2011). In addition, the taxation system of China is very complex, featured by massive tax incentives and multi-class legislations (Tang and Firth, 2011). The definition of reporting lag is adopted to explore the timing of annual earnings announcements

(8)

(Chambers and Penman, 1984). I expect the BTDs to be positively related to the delay, which means that higher tax avoidance leads to longer reporting lag.

Additionally, I examine whether audit quality affects the relationship of tax avoidance and the delay in annual earnings announcements and I expect that the audit quality has a positive association with the relationship between tax avoidance and the delay of the annual earnings announcement. Auditing has a relationship with earnings announcements because high quality audit is needed to disclose the failure and errors which can do harm to stakeholders’ reputation (Tendeloo and Vanstraelen, 2008). It is required that companies listed in China should get their reports audited each year.

Auditors can mitigate the agency problem to some degree by providing reliable and independent information to make investors have faith in the information provided by managers (Hayes et al., 2014). The conflicts of interests in agency theory can also be mitigated by auditors monitoring managers’ behavior (Fama and Jensen, 1983).

Furthermore, the investigation by the audit committees and the external auditors can somehow mitigate tax avoidance (Kim et al., 2011). The improvement of the audit quality can relieve tax avoidance, and might further influence timing of the annual earnings announcement by decreasing the delay.

I expect that there is a positive relationship between tax avoidance and the delay in the annual earnings announcement. And I further expect audit quality to have a positive influence on the main relationship. My results show that higher tax avoidance significantly

(9)

leads to longer reporting lag, and there is a significant association between the audit quality and the relationship of tax avoidance and the reporting lag.

My study has several contributions. First of all, my study contributes to the tax avoidance literature. Prior literature has examined definitions, determinants, and measurement of tax avoidance (Dyreng et al., 2005), but seldom emphasizes the influences of tax avoidance. My study fills the gap because I study how tax avoidance influences the timing of the annual earnings announcement on the basis of the agency theory, which suggests that tax avoidance has a positive impact on the managerial opportunism. My findings show that tax avoidance can actually affect the timing of annual earnings announcements. Higher level of tax avoidance leads to longer reporting lag.

The second contribution of my study is to the literature of the timeliness of annual earnings announcements. The earlier literature has studied some determinants of timing of the annual earnings announcements, yet they do not pay enough attention to the tax. I further fill the gap by adding the impact of tax avoidance and quality of audit to prior studies such as Sengupta (2004) and Kothari (2001) by testing the book-tax difference and the audit fee.

Thirdly, my study is of great interest to regulators because it involves the issue of how to release the relevant information to the decision makers when it is still useful for the decision making. Regulators want to make sure the information for users on the market is effective. I discover the association of the tax avoidance and the reporting lag as well as

(10)

the association of the reporting lag and the audit quality, both of which can help the regulators in guiding the market to disclose the earnings information at the most appropriate time. For example, the Chinese Securities Regulatory Commission (CSRC) may think about the tax avoidance when making accounting standards on disclosure, and the level of the tax avoidance can be limited to make sure the reporting lag would not be too long.

Fourthly, my study contributes to shareholders because they care about the benefit and harm by tax avoidance. Shareholders want managers to act on their behalf to make profit larger but managers can sometimes hide bad news. And by using the findings of my study, shareholders can have a better understanding of management behavior on tax avoidance and monitor managers by have an eye on delay in annual earnings disclosure. Shareholders may think of managers’ tax strategy when the reporting lag is long and monitor whether managers take proper tax avoidance activities to increase the profit.

Fifthly, my study is of interest to managers. Managers have to make decisions on the tax strategy as well as the timing to disclose, regardless of for their own purposes or on behalf of shareholders. By using my findings, managers may make a tradeoff between high level of tax avoidance and shorter delay in annual earnings announcements. For example, managers want to make the profit higher, and they want to take high level of tax avoidance. But they should be careful about the delay in earnings disclosure and make sure that would not be too long, otherwise it can reflect managers’ control problems in tax.

(11)

Finally, my study contributes to accounting literature on China. Prior studies such as Tang and Firth (2011) and Campbell and Wang (2014) have not found the relationship between tax avoidance and the delay of annual earnings disclosure based on the Chinese settings. I set this relationship in China because China has a very important role in global market and there is an interest for many companies to explore investments and partnerships in China. In addition, the tax system in China is very different from that in the United States or the Europe. Moreover, the topics of tax avoidance as well as the reporting lag are of great importance in China because the tax law system is still not complete and the awareness is very weak (Ma and Li, 2011) and the disclosure framework in China is still not very mature (Li and Yuan, 2005). Therefore, studying the tax and reporting lag in China are very interesting to both investors and management. Managers in China can know better about conditions of the tax avoidance and reporting lag in China to make their own specific plans. Investors both in China and other countries can know the investment conditions in China by reading the annual earnings announcements. They can learn about firms’ real performance from analyzing the level of tax avoidance and the reporting lag and decide whether or not to make investments.

My study has some limitations. First, I use the A-share firms in China to do the test but A-share firms have no earnings forecasts therefore earnings surprise variable is beyond control in my model. Earning surprise may bias the influences in my study. According to Crabtree and Kubick (2014), earnings surprise is significantly negatively associated with the delay in disclosure. If I exclude the effect by earnings surprise my results might be

(12)

more significantly positive, which could strength my hypothesis.

Moreover, I only include manufacturing firms in the observations because Chinese tax system differs by industries. Although manufacturing is the most important industry in China, it does not represent the whole market. I expect that the industries could lead to different results. If the tax rates of another industry are higher, firms will undertake higher tax avoidance. Therefore, there could be an increase in the coefficient of relationship between tax avoidance and reporting lag.

This thesis proceeds as follows. Section 2 discusses the literature review and hypotheses development for the study. Section 3 is about the methodology which covers the sample, model, variables and test. The result is discussed in Section 4 and conclusion is made in Section 5.

(13)

2 Literature and Hypotheses

2.1 Literature Review

2.1.1 Conditions in China

Historical development of tax system in China

Ever since the 1970s, a lot of corporate taxation reforms have been taken by the government in China due to the rapid development and the increasing importance of accounting information.

First, an overhaul of corporate income tax was issued in 1994 in China. Regardless of ownership types, all domestic firms had to pay 33% corporate income tax. Also, in 1994 there was a reform on the agency that collect the tax. The reforms in 1994 resulted in small aggregate welfare gains and the reforms may be too high so further improvements were needed (Toh and Lin, 2005).

On December 16, 1997, for the purpose of decreasing the delay in information, the Chinese Securities Regulatory Commission (CSRC) issued a law that required Shenzhen and Shanghai stock exchanges, which are two main stock exchanges, to work together with the release of annual reports (CSRC Pronouncement No. 113, 1997). However, firms still could manipulate the timing of annual earnings announcement to some degree (Haw et al., 2006). The new system was informative in earnings announcement and even though

(14)

firms were forced to report earnings in time, ‘good news early and bad news late’ still exits (Zhu and Yang, 2006).

In 2008, CSRC issued a new tax law to adjust corporation income tax and the tax burden was more fair. It changed the corporations’ actual tax and liability. Specifically, the tax rate changed from 33% to 25%. Furthermore, both individual and corporate income tax threshold were raised (Wen, 2012).

Subsequently, China entered the World Trade Organization (WTO) for the purpose of broadening its capital market furthermore to the whole world. And following that, CSRC later issued a new accounting standard in 2011 (Haw et al. 2006). The new standard, learnt from IAS-12-Income tax, first took deferred tax into account. More attention was drawn on the awareness of tax law (Ma and Li, 2011).

According to Gao (2011), the tax system in China has experienced several major reforms. The categories of taxes have been changed from 37 to 22 and the tax system structure is becoming more and more mature. Nevertheless, there are some issues left to be solved in Chinese tax system. The tax structure is still not balanced and the tax burden is still unfair in China (Ma and Li, 2011). Moreover, the tax law system is still not complete and the awareness is very weak (Ma and Li, 2011). At the present, corporate income tax is a very essential category of tax in China.

(15)

Current Tax system in China

China has three main kinds of taxes, which are sales tax and addition, value added tax, and income tax. For foreign-funded companies, China used to provide lots of preferential policies in tax. However, after the 2011 tax reform, the treatments have been reduced dramatically.

Sales tax and addition changes from 3% to 20% with regards to the specific industry. Such tax is regarded as part of sales prices. And this kind of tax can be discharged somehow by the sellers according to the standards. Moreover, companies have to pay a consumption tax, which is also a large part of the whole tax system. It can vary from 1% to 56% (Campbell and Wang, 2014).

For domestic goods, firms have to pay a value added tax with a rate of 13%. For imported goods, the rate is 17% and for exported ones the rate is 0%. It is not a part of sales price. Customers have to pay for it. Therefore, we cannot find it in annual report (Campbell and Wang, 2014).

Corporate income tax is 25% on average. However, it is different in different industries and among different kinds of firms. The corporate income tax for eligible small firm is 20%, for high-tech firm is 15% (Campbell and Wang, 2014; China State Administration of Taxation publications).

According to Li (2007), there are several differences in tax system between China and the United States. Chinese main tax is turnover tax, which is a kind of indirect tax, while

(16)

in the United States, income tax is the main tax which is a kind of direct tax. The Chinese government reduces or remits taxes to adjust the tax structure while the United States prefers to reduce the income tax rate. Moreover, the tax income departments are different in these two countries. In details, for individual income tax, China has nine levels of excess progressive tax rate while the United States has one level of excess progressive tax rate. For companies, value added tax is the most significant category of tax in China while in the United States, the most essential one is corporate income tax. For inheritance tax, it has been a very important category of tax in the United States, while in China, it is only at the starting stage.

Disclosure Framework in China

In China, all firms have to use the 31st_{of December as the fiscal year end and the filing}

deadline is April 30th_{. Firms have to disclose their annual reports and information through}

three kinds of channels. The first is the China Securities Daily, which is a significant kind of newspaper in China. The second one is Shenzhen stock exchange. And the last one is Shanghai stock exchange (Haw et al. 2006). China has three kinds of shares, which are A-shares, B-A-shares, and H-shares.1

The Chinese disclosure framework is still not very mature. There are some reasons for this problem according to Li and Yuan (2005). First, the accounting standards in China

A-share means shares that use the currency RMB. B-share means the shares that are available to foreign investors and H-share means shares listed in Hong Kong.

(17)

are still not perfect, which leads to insufficient in information disclosure. Furthermore, China is developing very fast, and the disclosure framework needs to be updated constantly to fit the reality.

2.1.2 Agency theory

The basic assumption for my research question is that it is managers that make decisions about corporate tax avoidance. Therefore, it is important to put this study in an agency framework.

Generally speaking, agency theory focuses on control problems due to the separation of ownership and control. Management is referred to as ‘agent’, and managers want to obtain confidence and support from ‘principals’, such as bankers, stockholders and employees. The problem is the agency costs among the agents and the principals. Monitoring cost is taken into consideration, which means the cost for the principals to pay to monitor the agents (Hayes et al., 2014). In addition, there exists the bonding costs which are intended to ensure that the agent is not going to undertake adverse actions that are not beneficial to the principals (Hayes et al., 2014). Moreover, residual loss should be taken into consideration, which means that there are still some losses taken place although the prior two costs have already been paid (Hayes et al., 2014). Agency costs of equity arise with delegation of decision rights from the shareholders towards the CEO (Jensen and Meckling, 1976).

(18)

As shown in Figure 1 (Healy and Palepu, 2001), there is always the information asymmetry within the agency theory. Managers have private information, which means that they have an advantage over the principals about the company. For example, managers know more about the company’s real performance such as its ability to repay loans while bankers know little, and they know better about the real profit than the stockholders do. However, managers also need to gain confidence from the principals, which means that they want investors to put money into the company (Hayes et al., 2014).

Auditors can mitigate the agency problem to some degree by providing reliable and independent information to make investors have faith in the information provided by managers. Therefore, it is important for both managers and principals to hire independent auditors (Hayes et al., 2014).

2.1.3 Good news early and bad news late

Consistent with agency theory, prior literature such as Kothari et al. (2008) generally shows that there are different incentives for managers to disclose good news versus bad news to investors. There are incentives for management to disclose bad news early. When mentioning litigation risk, managers tend to disclose bad news fast (Baginski et al., 2002). Managers can quickly release bad news while it is close to the grant date of option to lower the options’ exercise price (Aboody and Kasznik, 2000).

(19)

for managers to want to hide bad news. Firstly, they may hope to negotiate the possible contracts or to affect the sales of planned stock before announcing the unexpected bad news (Chai and Tung, 2002). Secondly, according to Trueman (1990), additional time is required to undo the bad news via accruals management. In this way, managers can better provide the plan to reverse poor performance or respond to criticism. Thirdly, managers withhold bad news on purpose because they want the industry-wide bad news to be released first so that they can justify reputational and litigation costs (Trueman, 1990). Fourthly, there might be high stress degree that companies choose to delay the announcement of bad news (Chai and Tung, 2002). Last but not least, other factors can contribute to managers’ decisions of hiding the bad news, such as managers’ career concerns, employment opportunities, loss of employment, and proprietary reporting costs (Verrecchia 1983; Kim and Son, 2014).

In sum, there are indicators for fast as well as delayed release of bad news. Nonetheless, in this study, the delay of bad news is mentioned because of the settings in China prefer the delayed release of bad news. The delayed news is considered as less relevant because the value of the information declined (Chai and Tung, 2002). Therefore, unexpected delay of news can lead to lower information quality.

(20)

2.1.4 Timeliness of annual earnings announcements

Financial information is the basic source of information that provided for investors. Therefore, it is very essential to publish timely disclosure (Ika and Ghazali, 2012). More timely information leads to less uncertainty for investors to make decisions and lower the level of information asymmetry (Ashton et al., 1989). Previous literature shows several determinants of earnings release dates, such as trading volume, institutional ownership, block ownership, technology firm, and undergoing acquisitions (Sengupta, 2004).

Companies disclose annual information mandatorily and voluntarily. This study focuses on mandatory disclosures, particularly earnings announcements. Prior literature defines the delay in annual earnings disclosure as the reporting lag. This shows the days between fiscal year-end and the release date of annual earnings announcements.

It is shown that market gets increasing earnings information when it is closer to the financial reports disclosure date (Kothari, 2001). Therefore, how much effective information is provided by earnings announcements to participants in the market can be the function of the reporting lag (Sengupta, 2004). There exists the issue of how to release the relevant information to the decision makers when it is still useful for the decision making. Regulators want to make sure the information for users on the market is effective. (Ika and Ghazali, 2012). Efforts are made in order to offer financial information in a timelier way. In China, there is a demand for making reporting lag shorter, which is required by the participants in the market (Zhu and Yang, 2006).

(21)

With regards to firm-level, in order to make optimal disclosure strategy, firms want to make sure that the strategy is effective for earnings announcements. In this way, firms’ optimal disclosure time can be influenced by the costs and benefits of the disclosure. For instance, a cost can be a significant competitive disadvantage and a benefit can be an enhanced image of the firm (Chen and Mohan, 1994). As a result, managers should choose the best announcement date based on their views’ of potential benefits to disclose the earnings by an earlier time.

2.1.5 Tax Avoidance

The literature for corporate tax avoidance is still at its starting stage, but there are active discussions on it (Hanlon and Heitzman, 2010). The magnitude and the determinants of corporate tax avoidance have caused great interest and have been widely studied (Hanlon and Heitzman, 2010). However, there still remains the problem that no universally accepted definition has been set for tax avoidance, tax aggressiveness or tax shelter (Hanlon and Heitzman, 2010). Therefore, different people actually have different opinions on tax avoidance.

Tax evasion is always mentioned when studying tax avoidance. The biggest difference is that tax evasion is not legal while tax avoidance is totally legal (Dyreng et al., 2005). Avoiding taxes does not show anything wrong. Firms can take measures to save taxes. In this way, tax avoidance is defined as the reduction of explicit taxes (Hanlon and Heitzman,

(22)

2010). It is a broad one. It means that tax avoidance shows all transactions that can influence company’s explicit tax.

In this study, tax avoidance is examined under agency theory. Studies show that on the basis of agency theory which details the conflicts of interests between managers and the principals, the activities of tax avoidance can further promote managerial opportunism (Desai and Dharmapala, 2006). Managers undertake the tax avoidance activities to conceal bad news, which contributes to the investors’ misunderstanding of the performance of the firm (Kim et al., 2011). For example, in the case of Enron’s Project Steele, earnings are controlled by management to mislead the investors by using complicated tax shelters (Desai and Dharmapala, 2009a). Moreover, managers can also justify the tax avoidance not to be transparent by saying that it is mandatory to be complicated and confusing for they want to lower the risk of tax avoidance being detected (Kim et al., 2011). Furthermore, incentive compensation can lower the level of tax avoidance because it mitigates the conflicts of interests under the agency framework, especially in poor corporate governance firms (Crabtree and Kubick, 2014).

Tax avoidance strategies are made by managers. There are two main views on tax avoidance in previous studies. The first is that managers undertake activities of tax avoidance because they want to reach their own benefits (Kim et al., 2011). In this way, investors may get more valuable information and managers are tended to undertake tax avoidance activities. The other one is that tax avoidance shows the sensitive relationship between managers and investors which take managerial opportunism into consideration

(23)

(Desai and Dharmapala, 2006).

However, it should also be mentioned that tax avoidance can benefit investors as well. Firms tend to make tax planning to save tax because it is good for investors to be residual claimants (Mills 1996). Shareholders want managers to act on their behalf to make profit larger, and managers would choose to lower tax liabilities when the benefit exceeds the cost (Hanlon and Heitzman, 2010). This means that investors will provide management with incentives to take tax-saving activities since tax avoidance is worthwhile (Hanlon and Heitzman, 2010).

2.1.6 Audit Quality

When auditors are able to discover breaches which relies on specific accounting system and report these breaches, the quality of the auditing is high (DeAngelo, 1981a). Prior studies use reputation and the conflict of power to explain audit quality (Deis, Jr. and Giroux, 1992).

The definition requires auditors’ competence and independence (Gul et.al, 2011). In the definition, discovering a breach means that auditors utilize their technical capabilities to discover the errors which can reflect auditors’ independence (DeAngelo, 1981a). Therefore, it is known that auditors’ independence is the key to audit quality. Furthermore, auditors’ independence cannot be well obtained if there is a lack of information on technical issues (Deis, Jr. and Giroux, 1992). An approach to make sure the maintenance

(24)

of auditors’ independence is to engage with large audit firms (DeAngelo, 1981). It is because that large audit firms do well in keeping brand name reputations and protecting them when they audit in a high-quality way (Gul et.al, 2011).

Auditing has a relationship with earnings announcements because the users of the reports ask for information which is reliable while auditors are required to work for interests of stakeholders. High-quality audit is needed to disclose the failure and errors which can do harm to stakeholders’ reputation (Tendeloo and Vanstraelen, 2008).

Auditing is also associated with agency theory. According to Chen et al. (2011), auditing can be seen as a tool to monitor management performance. Management discretion can be monitored to mitigate the asymmetry of information, which can decrease the degree of tax avoidance conducted by managers. Furthermore, earnings management can be affected by the monitoring quality when higher incentive plan is offered to managers. It implies that the quality of audit can influence the timing of annual earnings announcements (Chen et al., 2011). Chinese listed companies are required to audit their annual reports. The managers’ behaviors can be monitored by auditors, which can further mitigate the conflicts of interests in the agency theory (Fama and Jensen, 1983). In addition, to lower the effect of the agency theory, high quality audit is needed to offer related information to assess the actual performance of managers. (Tendeloo and Vanstraelen, 2008).

(25)

2.2 Hypotheses development

According to previous studies, in China, good news of earnings can be disclosed earlier than bad ones and managers can postpone the time to disclose bad news in comparison to the time of releasing good news (Haw et al., 2003; Fischer & Begley, 1998). The postpone of earnings news can be affected by several factors, such as institutional ownership, trading volume, technology firm, undergoing acquisitions, and block ownership (Sengupta, 2004). The objectivity of this thesis is to add tax avoidance as an essential element to determine annual earnings announcements’ timeliness.

Earnings announcements’ timeliness can be affected by tax avoidance for some reasons. Firstly, the relationship between the increase of managerial high-powered incentives and company’s tax avoidance can be shown by Desai and Dharmapala (2006). It can be shown that incentive compensation will reduce the tax avoidance level through decreasing the conflicts of interest under agency theory. In fact, tax avoidance can promote management opportunism and earnings can be controlled by managers. Herein, the interrelationship between management opportunism and tax strategy can be shown (Desai and Dharmapala, 2006). This means that activities of tax avoidance are taken by managers to control the earnings subsequently. The managerial view of earnings announcements can be shown by Mohan and Chen (1994). And they suggest that management behavior can influence the delay in annual earnings disclosure. In this way, if managers want to manage earnings by undertaking tax avoidance activities, they need more time before the disclosure of earnings information.

(26)

Moreover, evidence on a positive relationship between financial reporting and aggressive tax can be offered by Frank et al. (2008). It can be suggested that firms can participate in aggressive tax reporting and financial behavior in the same period of reporting. This implies that tax avoidance activities can influence the information shown in annual earnings announcements. Therefore, if firms want to take higher level of tax avoidance activities, more time is needed to provide more financial information that is going to be reported in annual earnings announcements.

Furthermore, Crabtree and Kubick (2014) show that if the tax system is very complex, more time is needed before the annual disclosure. China has a very complicated tax system with 19 categories of taxes and 9 excess progressive tax rates (Li, 2007). If firms in China want to take high level of tax avoidance activities, they have to take the complex tax system into consideration first. This means that they need more time to make proper tax strategy before the annual earnings announcements.

The above studies suggest that corporate tax avoidance is related to the timing of annual earnings announcements. Specifically, higher level of tax avoidance can cause more delay in annual earnings announcements. This leads to my main hypothesis:

H1: The level of tax avoidance is positively associated with the level of delay in annual earnings announcements in China.

Auditing can influence the relationship in H1 for some reasons. Firstly, auditors can mitigate the agency problem to some degree by providing reliable and independent

(27)

information to make investors have faith in information provided by managers (Hayes et al., 2014). Therefore, investors have the incentives to require for more independent auditors in order to get better information on decision making. In addition, the mitigation of agency conflicts can lead to more timeliness of annual earnings announcements (Hayes et al., 2014). Furthermore, the investigation by the audit committees and the external auditors can somehow mitigate tax avoidance (Kim et al., 2011). The improvement of the audit quality can relieve tax avoidance, and further influence timing of the annual earnings announcement by decreasing the delay.

Secondly, because of the complicated tax system, auditors need to do extra scrutiny over annual report (Crabtree and Kubick, 2014). And the high quality auditing is beneficial to increase the quality of annual reports so that audit quality is associated with the degree of strictness of tax (Tendeloo and Vanstrelen, 2008). This means that if the tax system is very complex, auditors need more time to do their work in order to provide high-quality information.

These lead to my second hypothesis:

H2: Audit quality positively affects the relationship of tax avoidance and the timeliness of annual earnings announcements in China.

(28)

3 Methodology

3.1 Sample

To investigate the relationship between tax avoidance and the delay in annual earnings announcement, I use archival (database) research approach.

I collect information from the database CSMAR. CSMAR is a unique, comprehensive database of China stock returns. All Chinese listed companies have to disclose their information on Shanghai stock exchange or Shenzhen stock exchange and CSMAR covers all the information disclosed.

The sample consists of companies with A-shares on both Shanghai and Shenzhen stock exchanges from the period 2012 to 2015. Shanghai and Shenzhen stock exchanges are two main stock exchanges in China and they include all listed firms in China. Shanghai stock exchange is based in the city of Shanghai while Shenzhen stock exchange is based in the city of Shenzhen.

I choose the A-share companies because A-share is more commonly used in China. Disclosures for B-share and H-share are relatively tight and their auditing requirements are under the International Accounting Standards, while A-share firms follow the Chinese Accounting Standards. Because A-share firms don not have earnings forecasts, I drop out the earnings surprise element.

(29)

I limit the industry to manufacturing industry because China has a very complicated tax system. Different industry contains different methods to calculate corporate income tax. Manufacturing industry is the main industry in China which has a very important position in the economy. Therefore, this industry can show most of the problems and conditions in China.

I examine the time period between 2012 to 2015, i.e. a 3-year sample period. The new accounting standard was issued by CSRC in 2011 and may influence my result. Therefore, data before 2012 is beyond control. And I choose the period after that change.

[Insert Table 1 about here]

Table 1 shows the sample selection screens. The initial sample for my study consists of 6,438 firm-year observations.

For the data to be included in my sample, some data are eliminated:

1. I eliminate ST firms. Since 1998, China has had Special Treatment (ST) in stock market. I drop these observations because these firms are normally firms that have uncommon poor financial conditions with very high risk of investment.

2. I eliminate observations that disclosed after 30th_{of April, which is the statutory filing}

deadline in China because if firms disclose after the statutory filing deadline some other problems will occur, which is beyond control (Zhu, 2006).

(30)

3. I eliminate observations with incomplete data or firms that whose data cannot be collected in stock exchanges because these observations’ performance cannot be detected.

4. I eliminate firms whose audit fee cannot be found from database because audit fee is an important variable in my study.

5. I eliminate some abnormal data such as negative listed years, and leverage exceeds 1, which may result from errors in CSMAR or cannot be found in the report. I eliminate these data because I want to examine the common condition.

The final sample size consists of 2,132 firm-year observations.

3.2 Measurement of tax avoidance

Lots of methods can be adopted to test tax avoidance according to Heitzman and Hanlon (2010). One method is the long-term efficient tax ratio measure computed as the total cash income taxes paid by cash in a decade divided by total pre-tax income, and special items are not included. Furthermore, the discretionary measure of tax avoidance can be computed with the usage of abnormal tax differences via the regression total book-tax differences on total accruals. In addition, the methods include efficient book-tax ratio as calculated through the division some estimates of tax liability by a means of cash flow or before-tax profits.

(31)

any tax plan that defers tax and because of the complex structure of Chinese tax system this method cannot include all the tax avoidance’s influence (Tang and Firth, 2011). According to Dyreng et al. (2008), long-run effective tax rate cannot be used because under this method, real activities, avoidance activities and targeted tax are not clear and also implicit tax is not included. According to Frank et al. (2009), the discretionary measure of tax avoidance is based on ETR measure so it cannot reflect deferral strategies either, and it is very hard to capture the so-called ‘unexplained portion’.

Therefore, the tax differences are used to test the tax avoidance level. All book-tax differences incurred by book-tax avoidance are included in the approach. BTDs is defined by Hanlon et al. (2005) as below:

BTDs = pretax book income - estimated taxable income - minority interest in earnings = book income - taxable income

Taxable income = (current foreign tax expense + current tax expense) scaled by statutory tax rate - firm’s change in net operating loss (utilizing the basic tax rate 25% for majority of the Chinese companies)

Both temporary and permanent differences are included in this method. Therefore, it is common to use this measure to test tax avoidance.

(32)

3.3 Measurement of the timing of annual earnings announcements

The timeliness of annual earnings disclosure can be measured by measuring the delay of the announcements, which also means the reporting lag.

Rlag = annual earnings announcement date - date of fiscal year end

In China, the fiscal year-end is 31st_{December. Therefore, the delay means the number}

of the days between 31st_{December and the release date.}

3.4 Research Design

My hypotheses state that tax avoidance and audit quality are related to the delay in annual earnings announcements. I use the model of Crabtree and Kubick (2014) because their model includes the tax’s influence on the reporting lag. I exclude their variables such as UE (earnings surprise), FYE (Fiscal year end), and Industry.

I exclude UE which means earnings surprise. UE is beyond control because China A-share firms do not show analysts’ forecast in earnings and there is no earnings surprise for these firms.

I exclude FYE because it is used to show whether a firm use a fixed date as the fiscal year end. This variable is meaningless because all Chinese have calendar year-end, which is 31st_{of December.}

(33)

I also exclude industry variable because in China different industries have different tax system and each industry has their own complicated tax rate. It would be less comparable to include all industries since they will definitely show huge difference. Therefore, in my study I use one industry only which is manufacturing firms whose more detailed categories of industry within manufacturing firms cannot be collected in CSMAR.

I add TAX to show the level of tax avoidance. I add FEE as an indicator to show whether the firm has high audit quality.

I use audit fee as an indicator to show the quality of the auditing. Higher audit fee would lead to less tax avoidance and less reporting lag. According to Chen et al. (2011), auditing can be seen as a measure to monitor management performance. It can mitigate information asymmetry and influence the timing of annual earnings announcements.

Nevertheless, it is hard to assess the audit quality. According to Tang (2010), auditors’ independence and the quality can be affected by the audit fee. The U-shaped type relationship between audit quality and audit fee has been studied by Zhang (2016). The audit quality will be low while the fee is too high or too low. Otherwise, higher audit quality can be ascribed to higher audit fee. Zhang and Zhang (2009) consider that audit fee can be a sound indicator of audit quality in China. In this thesis, the variable FEE is added to present the quality of auditing.

Therefore, my empirical model for testing both H1 and H2 would be:

(34)

β6BIG4i,t + β7SIZEi,t + β8MTBi,t + β9LEVi,t + β10SHAREi,t + β11PPEi,t + β12AGEi,t + β13EXCHi,t + + γt + ε

(1)

The control variables included in the model are as follows:

Firms with greater loss are not willing to disclose news early (Ajinkya et al., 2004). Therefore, if there is higher level of loss, managers are more likely to take tax avoidance activities. Thus, I add LOSS in the model and I predict to see a positive relationship between loss and delay in annual earnings announcements.

LOSSi,t = 1 if there is a net loss for firm i in year t

Previous research also suggests auditing to be linked to the timing of annual earnings announcements. Whittred (1980) shows that qualified audit report leads to the delay in annual earnings announcement. Moreover, if the firms have lower accounting quality, the delay would be greater. Companies and auditors need to conduct negotiation about the opinion that delays the announcements for more time is required to give opinion on the company. According to Xia (2012), the quality of audit will be enhanced and the delay will be decreased, if Big 4 audits the firm. Thus, I expect a negative relationship between the reporting lag and BIG4 and a positive relationship between the reporting lag and OPIN.

OPINi,t = 0 if the firm received a unqualified audit opinion, otherwise 1

(35)

According to previous studies, the size of company relates to the disclosure approach. Thus, annual earnings announcements’ timing can be affected by the size of company (Bushee et al., 2003; Frankel et al., 1999). I assume that Bigger firms perform better and they are monitored at a higher level by both auditors and investors to provide timely information and undertake proper tax avoidance activities. Therefore, I choose SIZE to measure the relationship and expect the them to be negatively associated.

SIZEi,t = natural logarithm of firm i total assets in year i

Because market competitors can also use the information offered by managers, firms’ profit may be lower. Firms tend to postpone the annual announcements and not to release essential information utilized by competitors due to the costs (Sengupta, 2004). Proprietary costs lead to the phenomenon (Cheon and Bamber, 1998). Moreover, if firms regard the costs to be lower there will more delays in the earnings announcements (Sengupta, 2004). Moreover, greater MTB means firms’ stocks are more valuable, therefore, firms’ stocks are more popular to investors. In this way, firms can attract more investment and have more growth opportunities, which means that firms have less incentives to take high level of tax avoidance and they have confidence to disclose news early. Therefore, I add MTB to show the impact of proprietary costs of the firm in the market. And I expect a negative relation between MTB and the reporting lag.

MTBi,t = market value of equity / book value of equity

(36)

total labilities to total assets. Firms with high leverage level have more debts. According to Lang and Lundholm (1996), leverage relates to company’s capability of growth in the future. Since that firms with higher leverage may have poorer performance which leads to more tax avoidance and higher level of delay, I expect that leverage relates to the reporting lag in a positive way as firms with higher leverage may have longer reporting lag.

LEVi,t = total liabilities / total assets

SHARE means the ownership concentration ratio of a firm. Firms with lower ownership concentration may have higher firm value, because it mitigates the conflicts of interests in the agency framework (Chen et al., 2005). Therefore, firms with higher SHARE can show worse performance which motivate managers to undertake tax avoidance activities and withhold the news. In this way, I expect SHARE to be related to the reporting lag in annual earnings announcements in a positive way.

SHAREi,t = common share outstanding / common shareholders

PPE captures a firm’s capital intensity. Stickney and McGee (1982) shows that if there are more depreciable assets, more tax avoidance activities are expected to happen. Firms with low PPE proportion have better efficiency in using the funds. Therefore, I expect a positive relationship between PPE and the reporting lag in annual earnings announcements.

PPEi,t = gross PPE / total assets

(37)

conditions in long-term. It is shown that the longer the listed years, the less likely for firms to receive qualified opinion (Zhu, 2006). Therefore, I expect firms with longer listed years to perform better and do not need to take high level of tax avoidance or delay. I expect AGE is negatively associated with reporting lag in annual earnings announcements.

AGEi,t = the total years of the company to be listed

I add EXCH to control for different stock exchanges’ influence on the relationship. China has two main stock exchanges, including Shanghai stock exchange and Shenzhen stock exchange. Different stock exchanges disclose information in different ways and have different level of strictness of the disclosure standards. According to Zhu (2006), the selection of stock exchange may have influence on the degree of reporting lag.

EXCHi,t = 1 if firm is from Shenzhen stock exchange, otherwise 0

Time indicators γt , and the clustered standard errors ε are also included.

I winsorized all the variables at the 1% and 99% level. And I use natural logarithm of independent variables to make the variables more stable. And I further did the robust regression clustered by firms to test the hypothesis.

To confirm my first hypothesis, I expect a positive sign on β1 + β3, the coefficients of tax avoidance in regression (1). To confirm my second hypothesis, I expect a positive sign on β3, the coefficient of the interaction term of audit quality and tax avoidance in regression (1).

(38)

4 Results

4.1 Descriptive statistics

Table 2 shows the time distribution. Data is collected from 2012 to 2015. Data during this period is evenly distributed, and there is no time effect in my study.

Table 3 shows the descriptive statics.

The mean value of the dependent variable, RLAG, is 93.813, with the standard deviation of 19.576. This indicates that the observed firms may disclose related information 3 months later after the end of the fiscal year (December 31). The first independent variable is TAX, whose mean value and standard deviation are 17.167 and 0.471, respectively. The second independent variable is FEE, whose mean value and standard deviation are 13.396 and 0.521, respectively.

LOSS is a dummy variable with a mean of 0.0304, which implies that 3% of the firms is loss-making. OPIN, which is also a dummy variable, has a mean of 0.016. This shows that 1.6% of the firms received a qualified opinion. The mean for the dummy variable BIG4 is 0.013, which means that 1.3% of the firms are not audited by a big4 audit firm.

(39)

SIZE has a mean of 21.498 which shows that the size of manufacturing firms is at an average level. Firms have an average size at around 20 in all industries (Wen, 2012). MTB has a mean of 3.427, which shows the market value to book value is about 3 to 1. And this shows a medium level of the popularity of firms’ stocks. LEV has a mean of 0.353, which shows that 35% of the assets are financed by debt rather than equity. Creditors prefer a low LEV and the average proportion in China is around 40% to 60% (Zhu, 2006), therefore my sample firms show an average level of leverage that is proper. PPE has a mean of 0.233 which means 23.3% of the total assets are PPE. Firms with low proportion of PPE show more efficiency in using funds and my sample firms in this study show a good PPE proportion because in other industry the PPE ratio could be lower. SHARE has a mean of 0.582, which shows that firms’ top 10 shareholders are holding 58.2% of the total ownership. Firms with higher SHARE show higher ownership concentration and lower capacity in asset restructuring and the sample firms in this study show an intermediate level of ownership concentration while it is common in China to see SHARE exceeding 80% (Zhu, 2006). AGE has a mean of 7.106. It means that firms have an average listed years of 7 years, which is not a very long period. And actually the two stock exchanges were set up in 1990. This means that a lot of firms may drop the market. EXCH has a mean of 0.770. This shows that 77% of the firms choose to report on Shenzhen stock exchange, which suggests that the disclosure standards in Shanghai stock exchange are stricter than those in Shenzhen stock exchange.

(40)

Table 4 shows the correlation coefficients for the variables.

The table shows that TAX and RLAG are negatively correlated with each other at -0.021, but the significance is low. This shows that higher tax avoidance is associated with shorter reporting lag, which contradicts my expectation.

FEE and RLAG are significantly positively correlated with each other at the level of 0.092 and the significance is at a high level. This is in line with my expectation that audit fee, which is an indicator of audit quality, has a positive impact on the delay in annual earnings announcements.

LOSS is positively correlated with reporting lag with a low significance at 0.046, which implies that firms suffering from net loss might release related information later as expected. OPIN is positively related to the reporting lag with a high significance at 0.069 as expected, suggesting that the companies whose accounting quality is lower might postpone the information disclosure. BIG4 is not correlated with RLAG and SHARE is not correlated with RLAG either, which are not in line with my expectation. This means that whether firms are audited by big4 audit firms or whether firms have highly concentrated ownership are not related to the delay in annual earnings announcements. No relationship between big4 and reporting lag may result from lack of power. SIZE is positively associated with RLAG with a high significance at 0.088, which means that bigger firms tend to delay the annual earnings announcements, which is not in line with my

(41)

expectation. It is suggested that the bigger the size of manufacturing firms, the longer the reporting lag. MTB is negatively related to reporting lag with a low significance at -0.045 as expected, and this means that the more popular a firm’s stock, the shorter reporting lag and if there is lower proprietary cost the level of delay in annual earnings announcements would be lower. LEV is positively related with reporting lag with a low significance at 0.053 as expected, which means that the bigger proportion of debt financing, the higher level of the delay. PPE is negatively associated with reporting lag with a low significance at -0.053 which is not in line with my expectation. This means that the more PPE in total assets, the shorter reporting lag and the more efficiency in using the funds, the less delay in annual earnings announcements. But it could also be reasonable because higher PPE ration means less complex of the firms and this can lead to shorter reporting lag. AGE is positively related to reporting lag with a low significance at 0.043, showing if firms have a longer listed year, they will have higher level of delay, which is not in line with my expectation. But it could also be reasonable because firms with longer listed years have more experience with gambling the system. Therefore, they tend to release the annual earnings announcements late. And EXCH is negatively related to the reporting lag with a high significance at -0.067. This could be reasonable because information in Shanghai stock exchange may be faster than that in Shenzhen stock exchange.

(42)

4.2 Hypothesis tests

I tested the effectiveness of my regression. First, I did the VIF test, which means the variance inflation factor test, to see whether multi-collinearity exists in the model. If there is multi-collinearity, the influences from the regression would be incorrect. The VIF test for this model shows that the variables in the regression have a lower value than 2 for VIF. This means that there is no multi-collinearity in this study.

I performed the Shapiro-Wilk test to see whether the residuals are normally distributed. By using the log transformation, the residuals are normally distributed.

Table 6 shows the regression results for both H1 and H2. The table shows the relationship between reporting lag and tax avoidance and audit quality, and controlled by the control variables.

I use β1 + β3 to test my first hypothesis. The sign on the coefficient of the tax avoidance variable is negative and significant at the 10% level (β1 = -0.019, p = 0.072), which is the main effect. It indicates that there is a negative correlation between the reporting lag and the tax avoidance tested with BTDs when the audit fee is at an average level. This is not in line with my expectation. The sign of the interaction term is significant at 1% level (β3 = 0.046, p = 0.010). The joint impact of β1 + β3 shows a significant

(43)

positive association with the reporting lag (β1 + β3 = 1.61, p = 0.094, one-tailed). This means that tax avoidance is significantly positively associated with the delay in annual earnings announcements as expected. The higher level of tax avoidance leads to the longer reporting lag. It is an economically significant result, because the increase in the coefficient of TAX and TAX*FEE by 1% can cause extra 4 days of delay, showing that the changes in book-tax differences as well as the interaction term have a significant result in the changes in reporting lag.

I use β3 to test my second hypothesis. The significance level of the sign on the interaction term coefficient is 1% (β3 = 0.046, p = 0.010), indicating a positive correlation between the reporting lag and the joint influence of FEE and TAX. This further means that audit quality is positively related to the relationship of tax avoidance and reporting lag. The high quality audit has a positive impact on the tax avoidance’s influence on the reporting lag. Therefore, my second hypothesis is confirmed. The economic significance of the results is that an increase in the coefficient of the interaction term by 1% can lead to extra 6 more days of reporting lag, which means that the changes in the audit quality have a significant influence on the tax avoidance and the reporting lag.

Apart from the above, results show that the sign on the coefficient of audit quality variable which is the main effect, is positive and significant at the 1% level when the TAX is at an average level (β2 = 0.024, p = 0.008). This means that audit quality, which is measured by audit fee, has a positive association with the reporting lag. And the higher the quality of audit, the lower the degree of the reporting lag. The economic significance is

(44)

that the changes in the coefficient of TAX can lead to extra days of the delay, which means that the changes in the quality of audit lead to significant changes in the reporting lag.

As to control variables, the sign on the coefficient of LOSS is positive and has a significance level of 5% (β4 = 0.075, p = 0.030) as expected, indicating that the reporting lag would be longer if firms have a net loss during the reporting year. OPIN is positively associated with reporting lag and significant at the 5% level (β5 = 0.096, p = 0.042) as expected, which means that the lower accounting quality of the firms, the longer the reporting lag. The sign on the coefficient of SIZE is positive and significant at the 5% level (β6 = 0.026, p = 0.027), which is not my expectation and shows that the bigger the firm size the more delays in annual earnings announcements. This could be reasonable if big firms have more accounts to be controlled and need extra time before the disclosure. The sign on the coefficient of MTB is negative and significant at the 10% level (β8 = -0.007, p = 0.057) as expected. This means that the more popular the firms’ stock, the more proprietary costs and the more growth opportunities, contributes to shorter reporting lag as indicated by the correlation coefficients. The sign on the coefficient of PPE is negative and significant at the 1% level (β10 = -0.177, p = 0.001) as expected, which means that the more PPE in assets the less delay will happen. This is an important indicator especially in manufacturing firms to see whether there is idle fund in the firms. This indicates that if firms have a low PPE ratio, firms have more efficiency in using the funds, and firms tend to have longer reporting lags. And other variables, BIG4, LEV, SHARE, AGE, EXCH, have no significant relationship with the reporting lag. This means that whether firms are

(45)

audited by a big4 audit firm or whether firms’ assets are financed by debt rather than equity or whether firms have high level of ownership concentration or whether firms have a long listed status or whether firms are reported on Shenzhen stock exchange, have no influence on the reporting lag.

In sum, these results confirm my first hypothesis which is that there is a positive relationship between tax avoidance and the delay in the annual earnings announcement. The joint impact of TAX and the interaction term suggest that there is a significantly positive association between tax avoidance and the reporting lag. This means that the higher level of tax avoidance, the longer reporting lag. These results are in line with and Kim et al. (2011) and Desai and Dharmapala (2006).

I also obtain extra information from the test of the first hypothesis. There is a negative correlation between the reporting lag and the variable TAX measured by BTDs, indicating that when the BTDs are higher, the reporting lag is shorter.

Furthermore, there is a positive correlation between the interaction term and the reporting lag. It means that the audit quality has a significant influence on the association between tax avoidance and timing of the annual earnings report, in accordance with Chen et al. (2011) and Hayes et al. (2014). Therefore, my second hypothesis is confirmed.

Additionally, the results also prove the positive relationship between the audit quality and the reporting lag. In other words, when the audit quality is higher, the reporting lag is longer.

(46)

5 Conclusion

This study examines how the timing of annual earnings announcement is affected by tax avoidance in and whether audit quality has an influence on this relationship in China. More timely information leads to less uncertainty for investors to make decisions and lowers the level of information asymmetry (Ashton et al., 1989). There are indicators for fast as well as delayed release of bad news, however, this study is set in China, where exists the ‘good news early, bad news late’ phenomenon. Management want to withhold bad news relatively to good news because managers think that if the news is about unexpected earnings which is negative, the bad news would be even worse. Also, managers tend to hide bad news because they need more time to undo the bad news or they do not want to have a disadvantaged position in the industry (Trueman, 1990). Therefore, management behavior can impact the timeliness of annual earnings announcements.

Tax avoidance can benefit investors as well as do harm to them. However, this study is set in China, where tax avoidance has more negative impact on investors. Managers’ tax-saving activities are tested under the framework of agency theory. There are conflicts of interests between managers and investors. Prior studies suggest that tax avoidance activities promote managerial opportunism, therefore, managers have more incentives to take tax plan to hide bad news (Desai and Dharmapala, 2006).

I expect that there is a positive relationship between tax avoidance and the delay in the annual earnings announcement because tax avoidance activities can promote

(47)

managerial opportunism (Desai and Dharmapala, 2006). Moreover, it can be suggested by Frank et al. (2008) that firms can participate in aggressive tax reporting and financial behavior in the same period of reporting. This implies that if firms want to take higher level of tax avoidance activities, more time is needed to provide more financial information that is going to be reported in annual earnings announcements.

I use audit quality to further detect the main relationship. High quality auditing can disclose failure and errors and auditors has a role in monitoring managers’ behavior (Fama and Jensen, 1983). Therefore, the high quality auditing is beneficial to increase the quality of financial reports and mitigate conflicts of interests. By improving the quality of audit, the extent of tax avoidance can be lower, which may have an impact of the timing on annual earnings announcements and decrease the delay.

The results of my regression test proves the positive relationship between tax avoidance and the delay in the disclosure of the annual earnings announcements based on the single-tailed test significantly, and this is what I expect. Moreover, the results also show a significantly positive association between the audit quality and the relation of the reporting lag and the tax avoidance. In other words, the audit quality affects not only the tax avoidance but also the reporting lag. I have an additional finding that there is also a significantly positive association between the audit quality and the reporting lag. When the audit quality is higher, the delay in the release of the annual earnings announcements is longer.

(48)

My study has several contributions. First of all, my study contributes to the tax avoidance literature. Prior literature has examined definitions, determinants, and measurement of tax avoidance (Dyreng et al., 2005), but seldom emphasizes the influences of tax avoidance. My study fills the gap because I study how tax avoidance influences the timing of the annual earnings announcement on the basis of the agency theory, which suggests that tax avoidance has a positive impact on the managerial opportunism. My findings show that tax avoidance can actually affect the timing of annual earnings announcements. Higher level of tax avoidance leads to longer reporting lag.

The second contribution of my study is to the literature of the timeliness of annual earnings announcements. The earlier literature has studied some determinants of timing of the annual earnings announcements, yet they do not pay enough attention to the tax. I further fill the gap by adding the impact of tax avoidance and quality of audit to prior studies such as Sengupta (2004) and Kothari (2001) by testing the book-tax difference and the audit fee.

Thirdly, my study is of great interest to regulators because it involves the issue of how to release the relevant information to the decision makers when it is still useful for the decision making. Regulators want to make sure the information for users on the market is effective. I discover the association of the tax avoidance and the reporting lag as well as the association of the reporting lag and the audit quality, both of which can help the regulators in guiding the market to disclose the earnings information at the most appropriate time. For example, the Chinese Securities Regulatory Commission (CSRC) may

Tax avoidance and the timing of annual earnings announcements in China