The effects of audit quality and the value relevance of financial statements

(1)

MSc Accountancy & Control, variant Accountancy Faculty of Economic and Business, University of Amsterdam

Master Thesis:

The effects of Audit Quality and

the Value Relevance of

Financial Statements

Final Version

Name: Cheniva Sabajo

Student number: 6161073

Date: 24th of June 2014

Supervisor: dr. Bo Qin

Co-assessor: drs. Mario Schabus

(2)

1

Abstract

(3)

2

and audit quality. Financial information is useful when it is relevant and reliable. I research whether the objective of financial reporting, to provide information to stakeholders which is useful in their decision making process, is achieved. In this paper I conduct and empirical investigations of the effects of earnings announcements on security prices.

A BigN auditor is expected to conduct higher quality audits, which increases the accuracy and reliability of financial information. Currently there have been multiple events and accounting scandals that contradict this assumption. However, this paper is not concerned with whether a BigN audit firm is actually superior to the non-BigN auditor, but more so with the perceptions among financial statement users. In order to measure the reaction of financial statement users, the Earnings response coefficient is used. The literature review is based on the agency theory, stewardship theory and the decision usefulness theory to determine the expectations according to literature. The research question states: what is the effect of audit quality on the information content of financial reports?

I conducted an archival research, to examine these expectations. I found that a relatively small amount of companies had Big4 auditors during the period 2012-2013. The sample suggest that 32 out of 660 companies had a Big4 auditor during the selected period and that even less firms chose to switch from or towards a Big4/non-Big 4 auditor during the period. In accordance with previous research, I find that the companies that switched from a non-Big4 to a Big4 audit firm had a significantly higher ERC relative to other firms. The descriptive statistics also show that the audit related fees of Big4 auditors are significantly higher than those of non-Big4 auditors. The results of the regression analysis shows that, as expected, multiple control variables affect the earnings response of companies. This suggests that the earnings numbers and other related ratios are relevant to users when making their investment decisions. Therefore, the value relevance of earnings information is high.. However, the results do not show a significant audit quality effect. Therefore, I cannot conclude that the quality and choice of the auditor, as well as a switch from or towards a higher quality auditor, affects the earnings response.

Keywords: Audit quality, Auditor change/switch, Earnings response, Value relevance, Decision usefulness, Agency theory, Stewardship theory, Big4 Audit firm.

Samenvatting

(4)

3

kwaliteit van de accountant. Financiële informatie is nuttig wanneer het relevant en betrouwbaar is. Ik onderzoek of de algemene doelstelling van financiële verslaggeving, om een bijdrage te leveren aan de informatieverschaffing van belanghebbenden., wordt bereikt. Van een groot accountantskantoor wordt een hoge kwaliteit aan controlevaardigheden verwacht. Dit draagt bij aan de juistheid en getrouwheid van de gepresenteerde financiële informatie. Momenteel hebben er zich meerdere gebeurtenissen voorgedaan en zijn er schandalen voorgevallen die deze assumptie tegenspreken. Echter, in dit artikel wordt niet deze desbetreffende assumptie onderzocht, maar eerder, de perceptie hiervan door gebruikers van financiële verslaggeving. Om de percepties van gebruikers van financiële verslaggeving in kaart te brengen, werk ik met de Earnings response coëfficiënt. Het literatuuronderzoek toont theorieën zoals die van de agency problematiek, stewardship en decision usefulness om de verwachtingen van de resultaten vast te stellen op basis van de beschikbare literatuur. De onderzoeksvraag luidt: wat is het effect van de kwaliteit van de audit op de informatiewaarde van financiële verslagen?

Ik heb een archivaal onderzoek verricht, gebaseerd op beschikbare data, om de verwachting the onderzoeken. Uit het onderzoek blijkt dat een klein deel van de bedrijven uit de steekproef een groot accountantskantoor aanwerven in de desbetreffende periode van 2012-2013. Slechts 32 van de 660 bedrijven contracteerde een groot accountantskantoor in deze periode, en zelfs minder maakten daadwerkelijk een switch van een klein naar een groot accountantskantoor of andersom. Uit de resultaten blijkt eveneens dat bedrijven die switchten van een klein naar een groot kantoor een significant hogere ERC hadden dan andere bedrijven. Ook blijkt dat de audit gerelateerde vergoedingen significant hoger waren voor cliënten van grote accountantskantoren dan die van kleinere accountantskantoren. De resultaten tonen eveneens aan dat financiële informatie en gerelateerde ratio’s relevant zijn voor gebruikers van financiële verslagen bij het nemen van hun investeringsbeslissingen. Om die reden, kan ik concluderen dat financiële informatie waardevol en nuttig is. Ik kan echter niet concluderen dat de kwaliteit van de accountant een significant effect heeft op de earnings response coëfficiënt.

Kernwoorden: Audit quality, Auditor switch, Winst reactie, Waarde relevantie, Decission usefulness, Agency Theorie, Stewardship theorie, Big4 accountantskantoor.

(5)

4

The objective and general purpose of financial reporting is to provide financial information about the reporting entity that is useful to potential equity investors, lenders, and other creditors in making decisions in their capacity as capital providers (IASB, 2014). This objective of financial reporting indicates that accounting standards are based on decision usefulness of information. Accounting standards, such as IFRS and GAAP, were introduced to provide organizations with a conceptual framework that outlines the elements of financial statements, and gives guidelines with respect to recognition, measurement, presentation and disclosure aspects (IASB, 2014). Financial standards are based on decision usefulness. This paper examines the usefulness of the information in financial statements for users.

In this paper I research whether the objective of financial reporting, to provide information to stakeholders which is useful in their decision making process, is achieved. I specifically analyse the information content of financial reports and the effects of audit quality on the earnings response coefficient. I conduct an empirical investigation on the effect of earnings announcements on security prices. I examine audit quality to research whether the choice of an auditor is important for investors, and whether investors react differently to financial statement information dependent on the auditor engaged.

The remainder of the paper proceeds as follows. First, I will provide background information, describe the research question, offer the motivation and contribution to literature and discuss the research methodology, main analysis and findings. Secondly, I describe the theory, literature review and hypothesis. Third, I discuss the variables and sample. This is followed by the results. Lastly, I present a conclusion of the findings.

1.1 Background

Recent accounting scandals at prominent companies such as Enron, HealthSouth, and WorldCom have shaken the confidence of investors (Agrawal and Chadha, 2005). The revelations about the unreliability of reported earnings continue to grow, as evidences by an alarming increase in the frequency of earnings restatements in the last five years. These events have changed the perceived accuracy of earning reports and the quality of an external audit statement. Investors worldwide are more and more questioning the validity of financial reports and audit procedures. During the past decade, the structure of the market for public accounting services, particularly audit services, also has received scrutiny from regulators, practitioners and researchers. The focus has been on the audit firms’ size, especially the large shares of markets held by a subset of firms called BigN audit firms. Following these events,

(6)

5

some major changes have taken place following these scandals (Agrawal and Chadha, 2005). First, the nature of the audit industry has changed. More Big audit firms have either divested or publicly announced plans to divest their consulting businesses. Second, Arthur Andersen, formerly one of the Big5 audit firms has gone out of business. Third, in 2002 the Sarbaneses-Oxley act was implemented. This law imposes a number of corporate governance rules on all public companies and their auditors. Lastly, in 2003 the New York Stock Exchange (NYSE) and NASDAQ adopted an additional set of corporate governance rules. These multiple legislative and regulatory changes were adopted in response to the widespread outcry that followed the scandals. Considering all these changes and the current economic state one might wonder what the current value of financial information and an audit opinion is to investors.

Ball and Brown (1968) were the first to examine the relevance of accounting income. The paper includes a graphic examination of the association between unexpected earnings and unexpected security returns. Ball and brown (1968) assume that net income is interesting to investors and that the market will incorporate news quickly. Therefore, changes in security prices will reflect the flow of information to the market. Their findings indicate that ninety percent of information is anticipated by the market and received by other more timely sources and ten percent of information in financial reports is ‘real’ news.

Francis and Schipper (1999) show that there has been a decline in the relevance of earnings information and an increase in the relevance of the balance sheet and book value information. These results are consistent with other research examining the value relevance of financial information (e.g., Collins et al., 1997; Ely and Waymire, 1999; Lev and Zarowin, 1999). Hail (2013) also analyses whether financial reporting have lost or regained relevance over the last 30 years. According to his research the Income statement has become less relevant over the last 30 years.

Multiple academics have researched the association between earnings announcements and the abnormal stock reaction. The usefulness and timeliness of accounting information, particularly earnings have been researched by Amir and Lev (1996), Aboody and Lev (1998) and Lev and Zarowin (1999), to mention a few. This topic is interesting because some academics argue that the economy has shifted from being based on tangible assets and manufacturing, to being based on intangible assets, services and information, and that accounting has not kept up with these changes (Landsman and Maydew, 2002). Some studies on the relationship between earnings reports and security price behaviour provide evidence suggesting that a significant portion of the information in earnings reports is reflected in the

(7)

6

security price changes (Ball and Brown, 1968; Brown and Kennelly, 1972). Investors also obtain information through other more timely sources. This same, more timely, information is for instance used by market agents to forecast earnings prior to their public release.

Atiase (1985) researches whether there are significant systematic cross-sectional differences in the security price reactions to earnings announcements of firms which are associated with specific firm characteristics, that lead to different amounts of pre-disclosed information. The findings suggest that the degree of unexpected security price changes in response to earnings reports is inversely related to the capitalized value and size of firms. Waymire (1984) finds that earnings announcements that give ‘good news’ are associated with significant positive abnormal returns in the days surrounding the publication, and vice versa for ‘bad news’. Some evidence in literature suggests that the value-relevance of earnings has declined over time (Collins et al., 1997; Francis and Schipper, 1999). Similarly, Hail (2013) writes:

“The analysis provides two general insights: first, the loss in relevance of the income statement, already documented in prior studies, continues over a more recent period and is not just limited to a select few countries but present in a large international sample. If anything, the decline in explanatory power is even more pronounced in countries with a strong institutional background. Thus, recent changes in the accounting standards, but also in the nature of the firms as well as the general economic environment like increased globalisation and market integration or the emergence of new technologies and alternative ways to disseminate information seem to have rendered reported earnings numbers less useful for valuation purposes”.(Hail, 2013)

Auditing reduces information asymmetries between managers and stakeholders by allowing outsiders to verify the validity of financial statements (Becker et al., 1998). Becker et al. (1998) research the relationship between audit quality and earnings management. There hypothesis indicates that there should be a negative relation between earnings management and audit quality. Because high quality auditors are more likely to detect questionable accounting practices, and when detected, to object or report to the earnings management attempts. Financial statements are useful to their users when they present fairly the economic performance and reality of the entity. Therefore, when earnings are managed (read manipulated) by executives, the information value of financial statements declines. Earnings management may lead (possible) investors, lender and other creditors to take different decisions or change their beliefs based on the information presented, than they otherwise

(8)

7

would. The earnings quality then declines, which decreases the information content of the reports. Thus, if the quality of the auditor is high, the amount of earnings management is expected to be low. Therefore, we can expect that financial reports that are audited by high quality auditors will have a higher information content and value relevance for investors. DeAngelo (1981) demonstrates that larger audit firms have greater incentives to detect and reveal management misreporting. Big6 audit firms are used to measure the audit quality. These audit firms are larger than their competitors, so it follows that they are of higher quality. One of the BigN audit firms write in their annual audit quality report:

“This report shows the actions we have taken, and continue to take, in support of our commitment to perform high-quality audits on a sustained basis, which is our highest priority”. (PWC, 2014)

Big4 audit firms endeavour to excel at their audits and the audit quality they deliver. The confidence of financial statement users could increase due to the use of a Big4 auditor. There is some empirical evidence that absolute and relative size of the audit firm serve as quality surrogates, particularly when a distinction is made between BigN and non-BigN firms (Palmrose, 1986). Previous research shows multiple trends concerning large audit firms: 1. the stock market reacts more favourably when a company switches to a large auditor rather than to a small auditor (Nichols and Smith, 1983), 2. large auditors give more accurate signals of financial distress in their audit opinions (Lennox, 1999), 3. companies with higher agency costs tend to hire large audit firms (Francis and Wilson, 1988), 4. large audit firms charge higher fees than small audit firms (Simunic and Stein, 1987), 5. companies involved in IPOs experience less under-pricing when they hire large audit firms (Palmrose, 1986), and 6. banks, investors and other financial institutions often require clients to hire large audit firms (Palmrose, 1986). This evidence shows that large audit firms are perceived to provide higher quality audits and provide greater credibility to financial statement users, relative to audits smaller size audit firms. Thus, large audit firms have reputations for performing high quality audits. In accordance with previously mentioned researches I also examine if auditor quality (based on their size) has an effect on the relationship of earnings announcements and security prices.

(9)

8 1.2 Research question

This research focuses on the value relevance of financial reports. I examine whether the financial report is useful to investors in their decision making process. Audit quality is included in order to determine whether audit quality has an effect on the perceived information content of financial reports. The research question states: What is the effect of audit quality on the information content of financial reports?

1.3 Motivation and Contribution

Recently there have been a number of scandals concerning accounting fraud. There are many examples of managers in organizations managing earnings or using real economic decisions to manipulate the earnings numbers. Earnings management occurs when managers use their judgement in financial reporting and in structuring transactions to alter financial reports, either to mislead some stakeholders about the underlying economic performance of the company or to influence contractual outcomes that depend on reported accounting numbers (Healy et al., 1999). The fact that managers manage earnings suggest: 1) they believe earnings do affect stock prices, and 2) they believe that investors will not see through their (opportunistic) behaviour.

This research has a contribution to current literature by determining whether earnings affect stock prices and whether audit quality has an effect on the information quality of financial reports. What distinguishes this paper from other literature is the research into the changes in shareholder confidence concerning the validity of financial reports and auditor reports. The primary distinction lies in the examination of value relevance gained or lost after an auditor change. Empirical research can help organizations to further increase the usefulness of financial reports by letting market response guide them as to what information is and is not valued by investors. As a result, companies can respond more efficiently to the information needs of investors. Which is beneficial to both the organizations, the investors and the stakeholders involved.

1.4 Main analysis and findings

A relatively small amount of companies in the sample had Big4 auditors during the period 2012-2013. The sample suggest that 32 out of 660 companies had a Big4 auditor during the selected period and that even less firms chose to switch from or towards a Big4/non-Big4

(10)

9

auditor during the period. In accordance with previous research, I find that the companies that switched from a non-Big4 to a Big4 audit firm had a significantly higher earnings response coefficient relative to other firms. The descriptive statistics also show that the audit related fees of Big4 auditors were significantly higher than those of non-Big4 auditors. The results of the regression analysis shows that, as expected, multiple control variables affect the earnings response of companies. This suggest that the earnings numbers and other related ratios are relevant to users when making their investment decisions. Therefore, the value relevance of earnings information is high. However, the results do not show a significant audit quality effect.

2. Theory, Literature review and hypothesis

2.1 Agency theory and Stewardship

An agency relationship arises between two or more parties when one, designated as the agent, acts for, on behalf of or as representative for the other, designated the principal, in a particular domain of decision-making problems (Ross, 1973). Essentially all contractual agreements in multiple settings between various parties contain elements of the agency theory. In organizational context the agent is the manager or executive who is responsible for the firm’s performance and the principal is the shareholder, who is the economic owner of the firm. The agency problem can be described as follows. Both the agent and the principal have independent utility functions and both act in a way to maximize their expected utility. By assumption the agent and principal have agreed upon a compensation scheme, to pay the agent for his service. The agent will choose actions that maximize his expected pay-off (Ross, 1973).

Essentially the principal (read shareholder) hires an agent (read manager) to perform specific duties that are in the principal’s best interest. However, when there is an incongruence of goals between the two, the agent may choose to pursue his own goals (Murphy and Jensen, 1990). In this case, the actions of the agent may not be in the best interest of the principal and be disadvantageous and costly.

Due to the asymmetry of information, the shareholder does not have perfect information regarding the executives’ activities and actions, and whether these are increasing or decreasing his wealth. To mitigate the inefficiency caused by the information asymmetry, the principal and agent agree that the agent will communicate ‘inside’ information on the

(11)

10

business performance to the principal (Baiman et al., 1987). This inside information is shown in financial reports. The information asymmetry could be reduced even further by hiring an independent auditor who is responsible for attesting the validity of the agent’s information. The primary role of the auditor is to produce stewardship information, that is, information used by the principal and agent for contracting purposes (Antle, 1982). Investors usually delegate decision making to managers. Afterwards they require information about the actions taken, for the purpose of controlling the managers. This links directly into stewardship (Gjesdal, 1981). An objective of financial statements is to report on the control and use of resources by those accountable for their control and use to whom they are accountable. This statement may serve as a definition of stewardship (Gjesdal, 1981).

The theory of the principal-agent problem and stewardship both explain the role of the auditor and the necessity for an external independent monitor who, provides shareholders with reasonable assurance concerning the validity of financial reports.

2.2 Decision usefulness theory

The decision usefulness approach has 2 perspectives: the measurement perspective and the information perspective (Scott, 2011). The measurement perspective indicates that managers are responsible for incorporating fair values in financial reports to assist investors in predicting firm values. The information perspective indicates that information is useful if it leads investors to change their beliefs and actions. The degree of usefulness can be measures by the price change following the release of information (Scott, 2011). In accordance with the theory it is possible to examine whether the information content in a financial reports is useful by analyzing the share price changes following its release. The information perspective concentrates on providing useful information within financial reports, because investors themselves are responsible for predicting future firm performance. This perspective assumes that security markets are efficient. Efficient security markets are markets in which prices fully reflect all available information. Thus, markets where the prices of securities at all time fully reflect all information that is publically known about those securities. Market prices would therefore adjust quickly to new information. On average, the total of all investor estimates will be unbiased according to the efficient security market approach. This approach recognized that the markets will react to useful information from any source, including financial reports.

The information approach links usefulness of the financial report to the information content of the financial report. According to this approach empirical accounting research adds

(12)

11

value by helping organizations’ management to further increase the usefulness of their financial reports, by letting market reactions guide them in determining what information is desired by investors.

The decision usefulness theory explains why we predict that the market price of a firm’s shares will respond to its financial report. The theory shows that investors have prior beliefs and expectations of future firm performance. After obtaining useful information from current financial reports, the investors revise their prior beliefs and expectations. This leads to a process of buying and selling securities, which causes share prices to fluctuate. Therefore, we can conclude that a change in share price reflects a change in investors’ beliefs and expectations of future firm performance. The stock price fluctuations are related to the earnings surprise. The earnings surprise reflects the difference between the actual earnings and the expected earnings. With the earnings response coefficient we can calculate the percentage change in stock prices after a dollar of unexpected earnings is reported. Thus, the earnings response coefficient shows the security price reaction related to unexpected earnings (Scott, 2011).

The decision usefulness theory aids in explaining why we have financial reports and why investors find them relevant, with regard to their decision making process. How relevant the information reported is to the market, can be measured by the earnings response coefficient.

2.3 Interaction Audit quality and Information relevance

In the previous section of the literature review I have explained why external auditors’ monitoring function exists and is useful, by using the theory of the principal-agent problem and stewardship. I have also explained why financial statements are relevant and how we can measure there relevance, using the decision usefulness approach. In this section I aim to connect the audit aspect with the information aspect.

An audit is defined as an unbiased examination and evaluation of the financial statements of an organization (Dictionary, 2014). DeAngelo (1981) defines the production of audits in term of inputs and outputs supplied by the auditor. Audit output can be characterized as independent verification of management-prepared financial data, and consists of a stated opinion with an associated quality dimension. The quality of audit services is defined to be the market-assessed joint probability that a given auditor will discover a breach in the client’s accounting system and will report the breach.

(13)

12

Auditing reduces information asymmetries between managers and stakeholders by allowing outsiders to verify the validity of financial statements (Becker et al., 1998). Becker et al. (1998) research the relationship between audit quality and earnings management. The hypothesis indicates that there should be a negative relation between earnings management and audit quality. His findings confirm these expectations. High quality auditors are more likely to detect questionable accounting practices, and when detected, to object or report on the earnings management attempts. Thus, if the quality of the audit(or) is high, the amount of earnings that are managed are expected to be low. DeAngelo (1981) shows that the larger the auditor measured by the number of clients, the less incentive the auditor has to behave opportunistically and the higher the perceived quality of the audit. Thus, larger audit firms have greater incentives to detect and reveal management misreporting. This could influence the value of audit quality determination by investors. In accordance with previous research I expect investors to perceive that larger audit firms conduct more thorough audits and provide them with a higher audit quality. Therefore, I expect the earnings response coefficient to be higher for client firms that are audited by BigN auditors versus client firms that are audited by non-BigN auditors. Teoh and Wong (1993) research the perceived audit quality and the earnings response coefficient. Their research shows that the earnings response coefficients of Big Eights audit clients are statistically significantly higher than for non-Big Eight clients. The contribution of this paper is found in the different research design. In this paper I research whether changes in the earnings response are associated with the switch from or towards a BigN/non-BigN auditor. Thus, I am examining the market reaction of an auditor switch.

Financial reports are defined as official public statements of a firm’s performance for a specific time period (Dictionary 2, 2014). The information value of financial statements is measured by the earnings response coefficient. The earnings response coefficient is a measure of earnings quality. Earnings quality measures, such as the accruals quality and earnings persistence, measure the quality of reported earnings, based on the fundamental qualitative characteristics: relevance and reliability. The earnings response coefficient is used in this paper, because, unlike the other earnings measures, it includes the market response on information provision. The information content of financial reports is an issue of obvious importance and is a focal point for many measurement controversies in accounting (Beaver, 1968). As Beaver (1968) paper, this paper also focuses on the reaction of investors after earnings announcements, which is reflected in the volume and price movements of common stocks in the weeks surrounding the announcement dates. In this research I have narrowed the scope by only including the abnormal stock return aspect and disregarding the abnormal stock

(14)

13

volume reaction. Additionally, I aim to determine whether the relationship between earnings announcements and stock price changes are associated with auditor quality by looking at the effects of an auditor switch.

To summarize, there is an information asymmetry between the managers of a firm and the investors (owners of the firm). Management has inside information on the firms’ performance and future performance expectations. A financial report should contain information on past performance and give an indication about future performance expectations. A manager could use the financial report to provide this inside information to investors. However, investors may be concerned about the accuracy and validity of the information presented. This concern may be reflected in shares being undervalued (Datar et al., 1991). The manager could hire an external auditor as a signalling device towards investors. Datar et al. (1991) demonstrates that the choice of the auditor and the resulting auditor report provides information to investors about the managers inside information and has a signalling effect. This shows that the value of a financial reports is affected by the auditor choice. If investors do react differently to earnings announcements due to the auditor choices, we can establish that auditor choice is relevant and has an effect on the information value of financial reports.

As stated in the previous sections, this paper aims to examine the effect of audit quality on the information content of financial reports. BigN audit firms are expected to deliver a higher quality auditor report. The audit opinion could then function as a signaling device toward investors. This provides more certainty on the accuracy and validity of the information in financial reports. However, recent accounting scandals, where (big) audit firms failed to detect and/or report on accounting errors, may have affected the contribution of an audit opinion, by investors. The hypotheses are formulated as follows: Hypothesis 1: The information in financial reports is relevant to investors. Hypothesis 2: Financial reports of big audit firm clients have a higher information quality than non-big/n audit firm clients.

(15)

14

H1: (+)

Monitor

Additional Assurance

H2: (+)

The firm represents the entity that provides financial information and is being audited. The auditor functions as a monitor of the actions by top executives and provides additional assurance to the investors of the entity. Two main problems/theories can be used in this model. The agency problem, which relates to an information asymmetry between the owners (investors) and managers (executives) of organizations, in which investors are not completely aware of firm performance. The stewardship theory suggest that investors may not be aware of whether the executives are acting in their best interest. Executives may hire auditors to signal to investors that there are no material errors in the financial statements and that they are making decisions that are in the best interest of the firm. To solve the agency problem, executives could provide private information in financial reports which enlightens investors on the firm performance and investment decisions. To solve the stewardship problem, audit firms may be hired to determine the quality of financial data. If the financial statement is relevant to investors, a positive effect to the earnings response coefficient could be expected. If the auditor report if relevant, I also expect a positive effect to the earnings response coefficient.

3. Sample, Variables and Empirical Model

3.1 Empirical model and Sample composition

The research conducted is an empirical archival examination. Data of US listed companies will be included for the year 2012 and 2013. This period is chosen because it provides the most current available information. The dataset will be obtained via the WRDS integrated database system, using the Audit Analytics database. The measurement of the theoretical

Investor Auditor Firm

Agency problemInformation Asymmetry Stewardship  Signaling

Financial statement (Private information on Firm performance)

Auditor report (Quality assertion of financial data)

(16)

15

concepts is done by utilizing previous research methods and models. These are described below.

To analyze the first hypothesis, I conduct an examination as done in the Ball and Brown (1968) paper. The information content of earnings is measured using the earnings response coefficient. This measures the association between unexpected earnings and abnormal stock returns (Ball and Brown, 1968). In accordance with the decision usefulness theory, the stock price reaction reflects the usefulness of information. Audit quality is associated with the size of the audit firm. DeAngelo (1981) shows that the larger the auditor measured by the number of clients, the less incentive the auditor has to behave opportunistically and the higher the quality of the audit. Consistent with the paper of Becker et al. (1998), I link BigN audit firms to (relatively) high audit quality and non-BigN audit firms are linked to (relatively) low audit quality.

To measure Audit quality  2 Dummy variables to measure the switch

toward/from Big4/non-Big4 audit firms (1,0)

To measure the information content of financial reports

 Earnings response coefficient (Association between unexpected earnings and abnormal stock return)

The empirical model is as follows

The sample consists of 660 companies which are active in various sectors/industries. The data is obtained using the Audit Analytics database (current auditors section). For all companies I obtained their financial, stock related and auditor information for the year 2012 and 2013.

In table 1, I show the data for the sample composition. Initially I obtained data of 1514 companies. After excluding companies that did not provide information on the auditor

(17)

16

engagement, the financial statement data or stock price data and those that did not have information for two consecutive years in the period 2012-2013, I was left with 660 firms. These firms represent the completed sample after the eliminations. Table 1 also reflects on the various industries in which the 660 companies is this sample are active. To obtain firm industry information, the SIC codes were utilized. The service, retail and mining industries are most represented within the sample and the agriculture, construction and transport industry the least. Furthermore, this research follows the difference in difference design, used in Ge, Tanlu and Zhang (2014). This research design entails that the firm performance of non-Big4 clients (control group) is compared with the firms that switched from or toward a non-Big4 audit firm.

TABLE 1

Sample, Industry composition and Variable definitions Panel A: Sample derivation

Description N

Initial sample from Audit Analytics 1514

Removal of firms with no Auditor information (379)

Removal of firms with missing financial statement data/stock price data (249)

Removal of firms not having two consecutive years of data (226)

Sample of firms for model in Auditor changes from 2012 to 2013 660

Panel B: Industry composition

Industry Description (Based on SIC) N N(%)

Agriculture 20 3% Mining 105 16% Construction 27 4% Manufacturing 64 10% Transportation, communication 27 4% Wholesale trade 65 10% Retail trade 144 22% Service 185 28%

(18)

17

No information 23 3%

Total 660 100%

This table shows the sample composition and industry distribution. 3.2 Dependent, Independent and Control Variables

The dependent variable is the earnings response coefficient, which measures the market reaction to earnings. The dummy variables used to measure audit quality are the independent variables. In this research I use ten control variables. These consist of the difference in total assets, the return on equity (ROE) and the change in the ROE, the return on assets (ROA) and the change in the ROA, the market-to-book ratio, the audit fees and the change in audit fees, the firm size and the fact whether a profit or a loss was reported. The variable definitions are listed in in panel C of Table 1.

TABLE 1 (Continued)

Panel C: Variable definitions

Variable Definition

Dependent Variable ERC An indication of the % ∆ in stock price after announcing 1$ of unexpected earnings. The ERC is measured by the stock price change divided by the amount of unexpected earnings.

Independent Variables Big4 Dummy Dummy which has a value of 1 if the entity has a Big4 auditor in 2012/2013, zero otherwise.

Dummy Switch-to-Big4 Dummy which has a value of 1 if the entity

has switched to a Big4 auditor, zero otherwise.

Dummy Switch-to-nonBig4 Dummy which has a value of 1 if the entity has switched to a non-Big4 auditor, zero otherwise.

(19)

18

Control Variables ∆ Total assets An indicator of the alteration between the amount of total assets in 2012-2013.

ROE An indicator of the return of equity (Net

income/Shareholder’s equity).

∆ ROE An indicator that shows the change in ROE

between 2012-2013.

ROA An indicator of the return on assets (Net

income/Total assets).

∆ ROA An indicator that shows the change in ROA

between 2012-2013.

Market-to-book ratio The market value of the firm divided by the

book value of the firm (market value reflects the share price multiplied by shares outstanding).

Audit Fees An indicator of the amount charged as audit

fees by the audit firm.

∆ Audit Fees An indicator of the alteration between audit

fees in 2012 and 2013.

Size An indicator of the size of the company,

established by using a continuous variable (total assets).

Profit Dummy which has the value of 1 when the

entity has positive earnings, and zero otherwise

The earnings response coefficient (hereafter ERC) is calculated by dividing the change in stock price by the amount of unexpected earnings. Unexpected earnings represent the actual earnings per share less the expected earnings per share. The actual earnings per share is calculated by dividing earnings minus dividend paid by the amount of shares outstanding. The expected earnings per share is equal to the actual earnings per share of the previous year

(20)

19

(Scott, 2011). The expected earnings per share is measured this way because I assume that investors are rational and rational investors would expect at least the same earnings per share in the following period, as they earned this period. To determine the percentage change in stock price, I subtract the closing price of two trading days after the earnings announcement by the closing price of the trading day after the earnings announcement. The trading price of the next trading day is utilized to prevent inequalities in situations where the earnings announcement takes place after trading hours (for instance). The method chosen to measure the ERC relates to the Ball and Brown (1968) paper and the Teoh and Wong (2006) paper. The ERC represents the relationship with a firm’s stock price and the information available to investors. This is the dependent variable. This entails that I try to establish whether the other variables listed have a significant statistical association with the earnings response.

The three independent variables show the firms that have engaged a Big4 auditor in the period of 2012-2013 or whether the firms have switched from or toward a Big4 audit firm. In order to establish if audit quality is associated with the earnings response coefficient, we have assumed that Big4 auditors perform higher quality audit related to their non-Big4 competitors. Therefore, I can determine whether the Big4 auditor clients have significantly different results, and whether a switch from or towards a Big4 audit firm is influencing the earnings response. The variable ∆ Total assets shows the change in assets for the company, during the period 2012-2013. This variable is used as a control variable, because the earnings response could also be dependent on the change in assets. If a company has increased their assets, investors may view this as a positive sign of firm performance. Therefore, a change is assets could be associated with the earnings response.

The variable Return on equity is calculated by dividing the net income by shareholder’s equity. The variable ∆ Return on equity shows the change in the ROE in 2012 and 2013. For investors the ROE may be viewed as an indicator of firm performance. Therefore, these variables may influence the earnings response, and are incorporated as control variables in this paper.

The Return on assets and the ∆ Return on assets represents, the net income dividend by the amount of total assets and the change of the ROA in 2012 and 2013 respectively. As the ROE, the ROA is a variable that may influence investors actions or beliefs when deciding to invest or not to invest in a potential firm. Therefore, this variable may be interestingly associated to the earnings response.

The Market-to-book ratio is used to determine the value of the company, by comparing the book value of a firms to its market value. The book value is calculated using

(21)

20

the firms accounting value on the balance sheet. While the market value is determined by the stock price changes, also called market capitalization. Investors use this ratio’s to determine the firm value. For the market-to-book ratio, the firm’s market value is divided by the book value. Because this variable also includes the firm accounting numbers and market value data, this is interesting to use in this research. The variables also may affect investors beliefs and action in their role as (potential) investor in the company.

The audit fees are the amounts charged by the auditor for their audit related work. The change in audit fees shows the change in audit related fees from the period 2012 to 2013. The amount of audit fees may also reflect the amount of auditor effort, to conduct the audit and provide the audit opinion. Therefore, investors may react to these variables.

The size of the firm may influence the relationship as well. Investors of larger firms may have different expectations of firm performance, dependent on the type and size of the firm they have invested in. Therefore, this variable is included as a control variable. The size of the firm is determined by utilizing the Total assets of the firm. This is a continuous measure that shows the magnitude of the firms property and effectiveness. The fact whether a company has been profitable or not may also affect the earnings response. Because, firms that are profitable may have higher abnormal returns.

These independent and control variables do not exclude for any other effects or competing explanations on the earnings response, but in they are very important variables to use in researching the relationship between earnings response and audit quality. The choice of which control variables to use was dependent on the papers by Becker et al. (2010), Teoh and Wong (2006) and Hoitash et al. (2012). Only those variables that could reasonably be expected to have an effect on the earnings response or auditor change/choice/quality were included. As in Hoitash et al. (2012), I also chose to incorporate the ratios themselves, as well as the change in the ratios during the period (e.g. ROA and ∆ROA). The idea behind this is that a rational investor looks at the current years financial and performance ratios, but also compares these to those of previous periods and make decisions accordingly.

(22)

21

4. Results

4.1 Descriptive Statistics TABLE 2

Summary Statistics

Panel A: Frequency of Audit firm switch in period 2012-2013

Description N N(%)

Firms with a Big4 Auditor during period 32

Firms that switched toward a Big 4 Auditor during period 5 1

Firms that switched toward a non-Big4 Auditor during period 12 2

Firms that did not switch during period 643 97

Total Companies 660 100

Panel B: Frequency by industry

Description Big4 Switch

Agriculture 0 0 Mining 3 2 Construction 3 0 Manufacturing 8 4 Transport, communication 1 1 Wholesale trade 4 2 Retail 2 1 Service 10 6 No information 1 1 Total 32 17

This table show the frequency of Auditor changes per category and industry as is manifested in the sample. Table 2 shows the descriptive statistics. The number of firms that had a Big4 auditor during the period 2012 and 2013 is 32. In this research, a total of 660 firms where included. Looking at the frequencies, it is shows that a small amount (N=5) of firms switched from a non-Big4 auditor to a Big4 auditor during the period (N=12). More firms chose to switch from a Big4 auditor to a non-Big4 auditor during this period. However, looking at the total sample, not

(23)

22

many firms made a switch between the different types of auditors. Agrawal and Chadha (2005) show that the Big4 audit firms handles 67 percent of audit engagements and collected over 94 percent of audit fees. They find that the audit market in many developed economies exhibits this concentration. However, in my sample, the Big4 auditors don not have a significant amount of engagements in the period 2012-2013. Also, the auditor switches in the sample show that more often firms switch from a non-Big4 auditor to a non-Big4. Also, firms that do chose to switch auditor, often switch to an audit firm with similar characteristics (from BigN auditor to BigN auditor and vice versa).

TABLE 3

Summary Statistics

Panel A: Descriptive statistics

Variables

non-Big4 Big4

(N=628) (N = 32)

Mean Median Mean Median

Difference in Mean P-value T-test (A) (B) (C) (D) (C) - (A) ERC (x1000) 14.94 0.22 18.62 0.11 3.68 0.793

∆ Total Assets (x10 million) 85.74 5.4 442.66 97.43 356.92* 0.023

ROE 2.27 0.80 2.80 1.74 0.53 0.430 ∆ ROE 1.82 0.52 2.17 1.09 0.35 0.515 ROA 0.09 0.05 0.08 0.06 0.01 0.642 ∆ ROA 2.18 0.07 0.30 0.07 -1.88*** 0.000 ∆ Market-to-book ratio 59.08 1.04 59.96 1.35 0.88 0.988 Audit fees (x1000) 61.31 35.00 254.93 200.55 193.62*** 0.000 ∆ Audit fees (x1000) 2.15 0.00 59.39 0.00 57.24 0.123 Firm Size 325.67 22.84 2080.88 582.29 1755.21*** 0.003 Profit 0.89 1.00 0.94 1.00 0.05 0.272 . ***, **, * indicates that the variables for the group Big4 are significantly different from the

companies that did not have Big4 auditors during the period at 1%, 5% and 10% sign. level

respectively, based on a two-tailed t-test for the mean. All the variables are defined in Table 1 Panel C.

(24)

23 TABLE 3 (Continued)

Summary Statistics

Panel B: Descriptive statistics

Variables non-Big4 Switch to Big4 (N=628) (N = 5)

Difference in Mean

P-value T-test

(A) (B) (E) (F) (E) - (A)

ERC (x1000) 14.94 0.22 37.30 28.20 22.36*** 0.000

∆ Total Assets (x10 million) 85.74 5.4 72.82 63.29 -12.92 0.275

ROE 2.27 0.80 3.43 2.28 1.16 0.501 ∆ ROE 1.82 0.52 3.19 1.53 1.37 0.493 ROA 0.09 0.05 0.10 0.10 0.01 0.565 ∆ ROA 2.18 0.07 0.09 0.07 -2.09*** 0.000 ∆ Market-to-book ratio 59.08 1.04 2.96 0.42 -56.12** 0.014 Audit fees (x1000) 61.31 35.00 190.64 89.10 129.33 0.312 ∆ Audit fees (x1000) 2.15 0.00 55.27 0.00 53.12 0.411 Firm Size 325.67 22.84 707.09 168.45 381.42 0.525 Profit 0.89 1.00 1.00 1.00 0.11*** 0.000

***, **, * indicates that the variables for the group switch to Big4 are significantly different from the companies that did not have Big4 auditors during the period at 1%, 5% and 10% sign. level respectively, based on a two-tailed t-test for the mean. All the variables are defined in Table 1 Panel C.

(25)

24 TABLE 3 (Continued)

Summary Statistics

Panel C: Descriptive statistics

Variables non-Big4 Switch to non-Big4 (N=628) (N = 12)

Difference in Mean P-value T-test (A) (B) (G) (H) (G) - (A) ERC (x1000) 14.94 0.22 41.11 0.14 26.17 0.479

∆ Total Assets (x10 million) 85.74 5.4 517.76 123.64 432.02 0.143

ROE 2.27 0.80 1.97 1.74 -0.3 0.589 ∆ ROE 1.82 0.52 2.30 0.48 0.48 0.585 ROA 0.09 0.05 0.04 0.03 -0.05*** 0.005 ∆ ROA 2.18 0.07 0.37 0.07 -1.81*** 0.000 ∆ Market-to-book ratio 59.08 1.04 39.36 2.46 -19.72 0.613 Audit fees (x1000) 61.31 35.00 216.41 232.50 155.10*** 0.000 ∆ Audit fees (x1000) 2.15 0.00 -18.86 0.00 -21.01 0.403 Firm Size 325.67 22.84 2313.87 1002.32 1988.2** 0.042 Profit 0.89 1.00 0.92 1.00 0.03 0.730

Table 3 Panels A, B and C show the descriptive statistics of the sample groups. The firms that had a Big4 auditor or made a switch from or towards a Big4 auditor are compared to the control group that did not have a Big4 auditor and did not switch. In order to obtain results from the sample that are representative for the population, I winsorized all continuous variables at the top and bottom 1% of the data. To exclude these ‘outliers’ I replaced the over extreme values with the most extreme, but acceptable, value at the top and bottom. The

***, **, * indicates that the variables for the group switch to non-Big4 are significantly different from the companies that did not have Big4 auditors during the period at 1%, 5% and 10% sign. level respectively, based on a two-tailed t-test for the mean. All the variables are defined in Table 1 Panel C.

(26)

25

complete summary statistics, for all the sample groups, are added to the appendix (shown in annex 1-4).

In table 3 panel A, I examine the characteristics of Big4 audit clients and non-Big4 audit clients. The mean represents the average of the numbers, while the median is the middle value in a list of numbers. After calculating the mean and median for Big4 and non-Big4 audit clients for multiple characteristics, I compared the means and conducted a two-tailed T-test for comparing two means using independent samples when the populations have unequal variances. This analysis shows that the difference between means is significant for the change in total assets, the change in return on assets, firm size as well as the audit fees. Based on these results I can conclude that Big4 audit clients have significantly higher total assets and significant changes in return on assets compared to non-Big 4 audit clients. The audit fees are also significantly higher for Big4 audit firms, compared to their non-Big4 competitors.

Panel B of table 3 shows the difference between non-Big4 audit clients characteristics and the clients that switched from a non-Big4 to a Big4 audit firm. The ERC, the ∆ ROA, MtB ratio and profit variable are significant. This entails that the earnings response is significantly difference for firms that switched from a non-Big4 to a Big4 auditor compared to firms that did not switch and did not have Big4 auditor during the period. This means that the responds to earnings information by Big4 clients leads to a significantly different change is stock prices, compared to non-Big4 clients. Thus, the decision usefulness of information for Big4 clients is significantly higher relative to firms that engage non-Big 4 auditors. The change in return on assets over the period is also different. The Big4 clients have significantly lower ∆ returns on assets in 2012-2013. This may be due to the higher audit related fees and other expenses that Big4 audit clients face relative to non-Big4 clients. Furthermore, the firms that switch to a Big4 auditor are significantly larger and have a larger profitability relative to firms that did not switch and did not have Big4 auditor during the period.

Panel C of table 3 shows the comparison between non-Big4 clients and the firms that switched from Big-4 auditors to non-Big4 auditors. These firms have significantly different ROA, ∆ ROA, firm size and audit fees. The significant decrease in return on assets and higher audit fees were not predicted effects. These results show that the ERC is not significantly different for firm that have or switch from/toward Big4/non-Big4 audit firms.

(27)

(28)

27

4.2 Test of Baseline predictions (Regression Analysis)

In order to examine the baseline predictions (hypothesis) as stated earlier in the paper, I use a regression analysis. The results from the regression analysis are shown in table 5. The actual regression outputs and additional information is displayed in the appendix. The regression analysis as conducted in table 5 for ERC1 is as follows.

ERC = β1 + β2 BIG4 + β3 SWITCH BIG4 + β4 SWITCH NON-BIG4 + β5 ∆ TOTAL ASSETS + β6 ROE + β7 ∆ ROE + β8 ROA + β9 ∆ ROA + β10 ∆ MtB ratio + β11 AUDIT FEES + β12 ∆ AUDIT FEES + β13 SIZE + β14 PROFIT + Ɛ

A regression analysis shows how one or multiple independent variables affect the dependent variable based on a significance interval (Alfa). This regression function shows how the dependent (Big4 dummies) and control variables affect the ERC. The first three variables are the dummies used to indicate whether a firm has a Big4 auditor or has switched from or towards a Big4/non-Big4 auditor. The additional variables are the control variables included. The results show that the Big4, Switch Big4 and Switch non-Big4 variables all have insignificant coefficients (Annex 5). According to the literature review, I expected positive coefficients for the Big4 and Switch to Big4 variables, but a negative coefficient for the Switch to non-Big4 variable. I could reasonably expect that when audit quality is high or increases (the company has or switched to a Big4 auditor), the earnings responds would have a significant positive value. When audit quality decreases (the company has switched to a non-Big4 auditor), I could reasonably expect the earnings response to be affected negatively. However, this regression analysis shows that the value of the coefficients is negative for BIG4 and SWITCH BIG4, but positive for SWITCH NON-BIG4. Also, neither one on the related P-Values is significant. The control variables ROA, MtB ratio, Audit fees, delta Audit fees and Size have significant P-values. Thus, these control variables significantly affect the earnings response. Surprisingly, return on assets has a negative coefficient. Suggesting a negative relation to the ERC.

The regression formula used in table 5 for ERC2 and ERC3 is as follows:

ERC = β1 + β2 ∆ TOTAL ASSETS + β3 ROE + β4 ∆ ROE + β5 ROA + β6 ∆ ROA + β7 ∆ MtB ratio + β8 AUDIT FEES + β9 ∆ AUDIT FEES + β10 SIZE + β11 PROFIT + Ɛ

In table 5 ERC2, I examine the regression of a sample, only including only Big4 audit firms. These are the firms that had a Big4 auditor during the period 2012-2013. Table 5 ERC2 shows

(29)

28

the regression result for the sample of firms that did not have a Big4 auditor nor switched from/towards to a Big4 auditor during the period. This sample could be named the control group. A comparison of the results from the regression analysis could shed more light on the differences between the firms in the two sample groups. Table 5 ERC2 shows that only the variable profit, is significantly associated with the ERC. The R square has a relatively high value of 0.60 compared to the R square in ERC1 and ERC3. Table 5 ERC3 shows multiple but somewhat different variables that are associated to the earnings response (ROA, ∆ ROA, ∆ MtB ratio, Audit fees, ∆Audit fees and Size).

TABLE 5

The determinants of the Earnings response coefficient

Total Sample Big4 non-Big 4

Independent Variables Dependent Variable

ERC (1) ERC (2) ERC (3)

∆ TOTAL ASSETS 0.00 0.00 0.00 ROE 254.70 2087.69 179.46 ∆ ROE -862.37 -848.37 -763.13 ROA -13554.90 -54663.10 -16087.08* ∆ ROA 1711.58*** -14711.99 1755.31*** ∆ MtB ratio 48.86*** 91.30 50.35*** AUDIT FEES 0.14*** 0.24 0.13*** ∆ AUDIT FEES -0.12** -0.23 -0.12** SIZE 0** 0.00 0* PROFIT 8649.71 189315.82*** -516.99 N 660 32 628 R Square 0.30 0.60 0.31 Adj R Square 0.29 0.42 0.30

This table reports the logistic regression summary output for the sample. ERC (1) shows the regression output for the total sample, ERC (2) for the firms that had a Big4 auditor and ERC (3) for the firms that did not have a Big4 auditor during 2012-2013. *,**,*** denote two-tailed statistical significance at 10%, 5% and 1% respectively. All variables are explained in Table 1 Panel C.

(30)

29

Overall the results reflect that, as expected, multiple of the control variables affect the earnings response of companies. This suggest that these earnings numbers and other related ratio’s are relevant to users when making their investment decisions. Therefore, the value relevance of earnings information is useful. Because this research only includes data from the period 2012-2013, I cannot state anything about the increase/decrease or fluctuations in the amount of usefulness provided by earnings information over time.

The regression results do not show a significant audit quality effect. As previously mentioned, in table 4, all the Big4 Dummy variables have an insignificant positive coefficient. Therefore, I cannot conclude that the quality and choice of the auditor, as well as a switch from or towards a higher quality auditor, affect the earnings response. Apparently this sample shows that investors view actual earnings information as more useful than the choice of the auditor and the audit opinion provided. Another limitation of these results is the fact that the sample group included a relatively small amount of firms that actually switched from or towards a Big4 firm within the period of 2012-2013. Because of this aspect in the sample group, the data analysis may over represent the determinant that affect the earnings response for non-Big4 clients that did not make any switches (the control group).

5. Additional Analysis

After analyzing the association between auditor changes and the earnings response, I now examine the determinants for auditor changes. Using a separate sample only providing auditor engagement information and the issues that lead to auditor changes. The sample consists of 1284 companies that switched from auditor during the period 2012-2013. Because no financial information is required for this sample, the sample obtained from the Audit Analytics database is much larger. The amount of companies in the sample that switched toward a Big4 auditor is 57 and the amount of companies that switched toward a non-Big4 auditor is 103. The rest of the firms in the sample, also called the control group, consists of firms that did switch from auditor during the period, but did not change between auditor types. This entails going from a Big4 auditor to another Big4 auditor or from a non-Big4 auditor to another non-Big4 auditor. As is shown in Table 6 Panel A, most of the firms in the sample that chose to switch, engage an audit firm that is of the same type as they previously contracted (N=1124). Furthermore, the results show that most firm that do change auditor type, move from a Big4 audit firm to a non-Big4 audit firm. In all sample groups, it is shown that going concern problems, bankruptcy or other disagreements leads to the termination of

(31)

30

the audit engagement in most cases. For the control group, internal control problems and accounting or restatement issues also occur often. This is followed by independence issues and scope limitations.

TABLE 6

Panel A: Determinants of Auditor Change

Control Group (N=1124) Switch Big4 (N=57) Switch Non-Big4 (N=103) Issue N N(%) N N(%) N N(%) Internal Controls 69 6% 4 7% 10 10% Accounting/Restatements 29 3% 3 5% 6 6% Independence/Scope limitations 21 2% 0 0% 4 4% Illegal acts 3 0% 0 0% 1 1% SEC investigation/interdiction 2 0% 0 0% 0 0% Fees Dispute 15 1% 0 0% 6 6%

Going concern or Bankruptcy 829 74% 40 70% 60 58%

(Other) Disagreement 156 14% 10 18% 16 16%

Total 1124 100% 57 100% 103 100%

This table shows the number and percentage of companies in the multiple sample groups that switched from auditor during 2012-2013. The control group refers to the companies that did not switch to a different type of auditor. The Switch Big4 group refers to the companies that switched from a non-Big4 auditor towards a Big4 auditor and vice versa for the Switch to non-Big4 group.

Table 6 Panel B shows the summary statistics for the additional analysis. The mean of the sample group and the difference in mean is displayed. The medians are not included because all the variables are Dummy variables with the values zero or one. Therefore these are the same values for the medians. The means of the sample groups are compared to the control group through a two-tailed T-test for comparing two means using independent samples. The results suggest that companies significantly switch from a Big4 auditor to a non-Big4 auditor due to issues with accounting or restatements, fees dispute and going concern or bankruptcy problems. Possibly because Big4 auditors have higher qualitative requirements for accounting

(32)

31

and strict guidelines for going concern opinions. Also, Big4 audit firms often have higher audit fees. The other results are not statistically significant.

TABLE 6 Summary statistics Panel B: Additional Analysis Control Group Switch to Big4 Switch to non-Big4

Variables Mean Mean

Difference in Mean P-Value T-Test Mean Difference in Mean P-Value T-Test

(A) (B) (B) - (A) (C) (D) (D) - (A) (E)

Internal Controls 0.061 0.070 0.009 0.788 0.097 0.036 0.158 Accounting/Restatements 0.026 0.053 0.027 0.224 0.058 0.032 0.059* Independence/Scope limitations 0.019 0.000 -0.019 0.298 0.039 0.020 0.166 Illegal acts 0.003 0.000 -0.003 0.696 0.010 0.007 0.231 SEC investigation/interdiction 0.002 0.000 -0.002 0.750 0.000 -0.002 0.669 Fees Dispute 0.013 0.000 -0.013 0.380 0.058 0.045 0.000*** Going concern or Bankruptcy 0.738 0.702 -0.036 0.550 0.583 -0.155 0.000*** (Other) Disagreement 0.139 0.175 0.037 0.438 0.155 0.017 0.644 This table shows the descriptive statistics for the additional analysis. The control group refers to the companies that did not switch to a different type of auditor. The Switch Big4 group refers to the companies that switched from a non-Big4 auditor towards a Big4 auditor and vice versa for the Switch to non-Big4 group.

(33)

32

6. Conclusion

In this paper I have examined the relationship between the value relevance of financial information and audit quality. Financial information is useful when it is relevant and reliable. The information could than assist financial statement users in making creditor/investment decisions. Therefore, I research whether the objective of financial reporting, to provide information to stakeholders which is useful in their decision making process, is achieved. I conduct an empirical investigation on the effects of earnings announcements on security prices.

A BigN auditor is expected to conduct higher quality audits, which increases the accuracy and reliability of financial information. However, recent accounting scandals, involving several BigN auditors may have raised the question whether BigN auditors actually perform higher quality audits, compared to their smaller competitors. However, this paper is not concerned with whether the BigN auditor is actually superior to the non-BigN auditor, but more so with the perceptions among financial statement users. In order to measure the reaction of financial statement users, the Earnings response coefficient is introduced. This variables show the change in stock price after unexpected earnings have been presented. In contrast to other earnings quality measures, the earnings response shows the factual market reaction. For that reason, I chose to work with the ERC. The literature review explained the agency theory, stewardship theory, the decision usefulness theory and the link to audit quality. These theories where used to establish the predictions of the results according to the literature. The research question stated: what is the effect of audit quality on the information content of financial reports? The hypothesis where as follows: Hypothesis 1: The information in financial reports is relevant to investors. Hypothesis 2: Financial reports of big audit firm clients have a higher information quality than non-big audit firm clients.

According to previous literature I could reasonably expect that larger audit firms would provide higher audit quality. This leads to higher quality financial reports and should increase the decision usefulness of financial statement users. This increase in usefulness is expected to be reflected in the Earnings response coefficient. While using a sample of 660 firms and conducting an archival empirical research, to examine these expectations, I found that a relatively small amount of companies had Big4 auditors during the period 2012-2013. This is remarkable because multiple sources would suggest a concentrated market for the

(34)

33

Big4 audit firms. The sample suggest that 32 out of 660 companies had a Big4 auditor during the selected period and that even less firms chose to switch from or towards a Big4/non-Big 4 auditor during the period. The results also suggest that most switched took place for firms active in the service and manufacturing industry. In accordance with previous research, the descriptive statistics show that the companies that switched from a non-Big4 to a Big4 audit firm had a significantly higher ERC relative to other firms. The descriptive statistics also show that the audit related fees of Big4 auditors are significantly higher than those of non-Big4 auditors. The results of the regression analysis shows that, as expected, multiple of the control variables affect the earnings response of companies. This suggest that the earnings numbers and other related ratio’s are relevant to users when making their investment decisions. Therefore, the earnings information is proven to be useful. The regression results do not show a significant audit quality effect. As previously mentioned, in table 5 (Annex 5), all the Big4 Dummy variables have an insignificant positive coefficient. Therefore, I cannot conclude that the quality and choice of the auditor, as well as a switch from or towards a higher quality auditor, affects the earnings response. Apparently this sample shows that investors view actual earnings information as more useful than the choice of the auditor and the auditor opinion provided.

Furthermore, the additional analysis shows that companies often switch from auditor but chose an auditor firm from the same type. Thus, going from a Big4 auditor to another Big4 auditor and vice versa for the non-Big4 audit clients. However, when changing auditor type, the results show that companies more often switch from a Big4 auditor to a non-Big4 auditor than from a non-Big4 auditor to a Big4 auditor. Considering the total sample and all switches made, the main reasons for changing audit firm are going concern or bankruptcy issues and disagreements. This is followed by internal control issues and accounting problems or restatement issues. The summary statistics of this section show that companies switch from a Big4 auditor to a non-Big4 auditor due to accounting or restatement issues, fees dispute and going concern or bankruptcy issues. This occurrence could reasonably be expected, considering both auditor type characteristics.

Some limitations to this research that may be taken into account in future research are: 1. Not including a stock volume effect, but only examining the stock price effect, 2. Using a relatively short and current time period (which may not reflect the economic reality in the long term) and 3. Having a sample with a relatively small amount of firms that switched from or towards a Big4 auditor (may cause some noise or biased results).

The effects of audit quality and the value relevance of financial statements

Master Thesis: