Adding additional information to the auditor’s report: The more, the better? Survey research among nonprofessional investors.

(1)

Adding additional information to the auditor’s report: The

more, the better?

Survey research among nonprofessional investors.

MSc. Accountancy

University of Groningen

Faculty of Economics and Business

January 2019

R. Bouwman

S.2582007 Gerard Doustraat 62 - 3 1072 VV Amsterdam +31 6 120 900 55 Rikbouwman@icloud.com

Supervisor: Prof. Dr. D.A. de Waard.

(2)

Abstract

This is an explorative research about the value of the extended auditor’s report from the perspective of nonprofessional investors located in the Netherlands. Using a survey, we analyzed whether the extended auditor’s report is being read and to what extent the annual report is being read. The goal of this study is to find out whether the investment goal, risk presence and education of the respondent influences the use of the extended auditor’s report as well as for the annual report. A sample of 118 respondents is used to run the statistical tests. This research adds to the current literature as it found various relations which have not been found before. First of all, the results indicate that nonprofessional investors are reading the extended auditor's report. A significant negative effect is found for the relation between risk preference and the use of the extended auditor's report, furthermore education is found as a positive significant moderating effect on this relation. The findings did support the expectation that education had a significant moderating effect on the relation between risk preference and the extent of reading the annual report, but education had no main effect itself. Whether a nonprofessional investor invests for retirement purposes or not, does not make any sense with regards to reading the extended auditors report or the annual report. This research might help regulators and accounting firms understand the benefits and points for improvement regards the extended auditor's report. It will also give investors an insight into the usefulness of the extended auditor’s report.

Keywords: Extended auditor’s report, Annual report, Nonprofessional investors, Investment goal, Risk preference, Education.

Acknowledgments:

I would like to thank some people for helping me during the time of writing my thesis. In particular, I would like to thank my supervisor, Prof. Dr. D.A. de Waard, for his support during the past half year. Dr. de Waard did not only provide me with useful feedback, but he was also thinking along with me during the whole process. He also managed the contact with the VEB and gave valuable input for the survey for which I am grateful. I want to thank Mr. J. Bonestroo, my supervisor from KPMG, for guiding me through KMPG as graduate and for his valuable comments towards my thesis. I would like to thank my two fellow students with who I developed the survey. And finally, I would like to thank all respondents for participating in my research. Without their input, this research could not have been executed.

(3)

“Inclusion of more information in the auditor’s report, like the main going concern

risks, the methodology, the reliability of the methodology and materiality

requirements. From an international perspective, standards are being extended

further so a more informative auditor’s report can be realized. I am in favor of

this.”

Jeroen Dijsselbloem

(4)

I. Introduction

1.1 Introduction

New York, 14 September 2008, the day that the New York Times announced the bankruptcy of investment bank Lehman Brothers. According to the journalists of the New York Times “one of the most dramatic days in Wall Street’s history,”1 a day where America went deeper into the financial crisis. The investment bank went bankrupt while they had an unqualified accountants’ opinion. The independent opinion should indicate that the financial statements are presented fairly and in accordance with the current standards of the Public Company Accounting Oversight Boards (“PCAOB”). The collapse of the financial sector raises questions by the public about the work performed and the value of the opinion of accountants (Sikka, 2009). There was a demand for change. Standards setters gave their reaction in the form a new format of the audit report where auditors have to provide more insight and information about the audit. The extended auditor’s report (hereafter: EAR) was born.

The demand for change did not come entirely unexpected. Various studies have examined the value of the former audit report in the past. Adding additional information to the audit report is already suggested in 1991 by Hatherly, Innes, and Brown (1991). Their study provided evidence that expanding the audit report can change the decision-making process of the readers (Hatherly et al., 1991). The work performed by auditors was a black box and only the final opinion became public (Church et al.,2008). Instead of reading the report, Gray et al. (2011) concluded that the intended readers do not read the audit report at all. “The average financial statement user is not interested in carefully evaluating the current audit report because it is such a standardized product” (Mock et al.,2013). When a company has an unqualified opinion, stakeholders assume that the audit was performed properly and they do not look further at the report. A consequence of the former audit report is that it maintains the audit expectation gap, “An audit expectation gap exists when there are differences in beliefs between auditors and the public about the duties and responsivities assumed by auditors and the message conveyed by the audit report” (Monroe and Woodliff, 1993). Without providing any company-specific information in the auditor’s report, stakeholders remain ignorant regarding the work performed by auditors. The demand for a change was also present in the Netherlands. This was among others observable in the yearly letters of the VEB (Vereniging van Effectenbezitters - Association of Dutch Investors)2 to the six most prominent accounting firms about the expected contribution of accountants for the coming year. Since 2013, the VEB writes letters to the accounting firms with suggestions and points of improvement regarding the information provided in the auditor’s report (VEB: 2013-2018). Simultaneously, the NBA (Koninklijke Nederlandse Beroepsorganisatie van Accountants - The Royal Netherlands Insitute of Chartered Accountants) came in 2013 with the announcement that

1_{Cited from an article of the New York Times, 14-09-2008:https://nytimes.com/2008/09/15/business/15lehman.html} 2_{The VEB is a Dutch association who represents the interests of nonprofessional investors.}

(6)

they would participate in a pilot of the International Auditing and Assurance Standards Board (IAASB), to provide more information about the audit in the audit report (NBA, 2013). After a positive evaluation of the pilot, the NBA introduced the new standard (Standard 702N) which obligates accountants to provide an EAR for audits of public listed companies for the fiscal year 2014 and the years following (NBA, 2014). Standards 702N adds three main topics to the former audit report, namely: 1) a materiality paragraph, 2) a paragraph with the key audit matters and 3) a paragraph with the scope of the audit. The new standard is set for a few years already, giving the auditors and stakeholders some time to get used to the new report. This makes it possible to measure whether the desired result is achieved. The EAR has two main functions, namely: 1) to enhance the knowledge and understanding of investors about the business and the work performed by the audits, and 2) more trust, confidence, and reliance in the work of an auditor (Pakaluk, 2017). These functions will, hopefully, enhance the transparency and reduce the expectation gap between society and accountants. Whether these goals have been achieved in the Netherlands remains a question.

1.2 Research question development

The audit report is pre-eminently the medium to communicate the work performed during the audit. The EAR can enhance communication with the relevant stakeholders and close the expectations and information gap (ICEAW, 2017). Previous research suggests that adding additional information to the audit report can increase the value of the audit (Mock et al., 2013; Harherly, Innes and Brown, 1991; Fakhfakh, 2015; Manson and Zaman, 2001; Pakaluk, 2017). According to Bedard et al. (2012), adding additional information especially for nonprofessional investors would be beneficial. This research, will focus on the relation between the nonprofessional investors and the extended auditor’s report.

Although investors ask for more information (VEB: 2013-2018) and researchers suggest that more information will add value, Gray et al. (2011) suggest that investors do not read the audit report at all. Instead of reading the audit report, investors prefer to rely on secondary information sources such as analysts’ recommendations even when the additional content about the audit has been added in the audit report (Gray et al., 2011; Coram et al.,2011). The information provided in the EAR should be included in the analysts’ recommendations, so there is an indirect value for investors, but even that is not the case. According to Coram et al. (2011), analysts do not read the audit report at all. Despite the new form of the audit report being in a very early stage in the time Coram et al. (2011) did their research, other studies provide the same suggestions. “The auditors’ disclosure reliably reflects companies’ financial reporting risks, but the disclosures have not significantly improved the information environment” (Lennox et al., 2018). Reason for this is that investors have other sources who indicate whether investing in a particular company is risky or not before the EAR was introduced. Critics of the new standard suggest that the EAR enhances the complexity of the audit report, which makes it less readable (PCAOB, 2016a). This could cause a hinder in the information adoption of the audit report instead of providing more information (Li, 2008). Therefore, some critics assume that investors do not use the EAR. However, Smith (2016)

(7)

concluded in his research that the new report form enhances the readability of the audit report and provides more client-specific financial risks. This is in line with the research of Reid (2018) who suggested that the EAR is enhanced informative from an investors perspective. The contradictory results of previous literature cannot provide us with an unambiguous conclusion. The evidence that the new audit report is informative supports the recommendations provided in the yearly letters of the VEB and Eumedion. However, we cannot just ignore the contradictory result of prior studies. This research will find out what the added value is of the EAR in the Netherlands. To what extent do the nonprofessional investors use the report in the Netherlands? This will be the main research question guiding this research.

RQ: Do nonprofessional investors, living in the Netherlands, use the extended auditor’s report as a useful information source in their decision-making process?

Besides the question of whether the EAR is being read at all, this research will focus on the personal preferences of nonprofessional investors namely: the goal of the investment and risk preference. This research will investigate whether there is a link between the investment goal of nonprofessional investors and the extent of using the EAR. A distinction will be made between retirement goals and non-retirement goals. From previous literature, the goal and investment horizon of investors influence their information search and evaluation process (Warren et al.,2014; Li et al.,2018). Secondly, this research will investigate the link between risk preference and the extent of using the EAR. The risk preference of nonprofessional investors influences the use and the search of information. Risk-averse and risk-tolerant investors act differently when it comes to information processing (Tseng and Yang, 2011). Furthermore, this study will examine whether education has an impact on the relation between risk preference and the extent of using the EAR. Educated individuals are better able to process complex information and are less likely to avoid this (Moshman, 1990; Chandra and Ravinder, 2011). It is expected that a moderating effect will occur because investors behold the financial statements and the accessory auditors report as complex. Since the EAR is provided in the annual report of the company, for reliability purposes, the relations with regards to the annual report will be tested as well.

1.3 Scientific contribution and practical relevance

This research contributes to the current literature as it explores the added value of the EAR as a whole for nonprofessional investors. Since the introduction of the extended auditors report, there has been some research to the usefulness, but the outcomes are not corresponding with each other. More research is crucial to be able to formulate a valid statement about the EAR. This research will examine the possibilities to find evidence that strengthens the research regarding the usefulness of the EAR.

Different regulators introduced the EAR: the UK in 2013, the Netherlands in 2014 and the US in 2017 (FRC, 2013; NBA, 2014; PCAOB, 2017). Since it is not an international uniform standard, there are differences in the extended auditors report among countries. Although the details may differ, the main idea behind the new standard is the same: to provide more information about the

(8)

audit. Therefore, this research will not make a distinction between the differences in standards in the literature review. A distinction between the countries will be made. The reaction of the public can differ among countries due to the legal and regulatory environment differences (Carver et al., 2017). This research contributes to the current literature regarding the usage of the EAR as it explores the reactions of nonprofessional investors in the Netherlands. The difference between what the society expects from accountants and what the accountant does is defined as the expectation gap in the literature (Liggio, 1974). Many studies suggested that better communication of accountants could help to close the gap (Kelly and Mohrweis, 1989; Gay and Schelluch, 1993; Monroe and Woodcliff, 1993). This research adds to the literature regarding closing the expectation gap since this is one of the goals of the EAR, and we will determine whether nonprofessional investors use the EAR.

This research adds value to the current literature since it focuses on the effectiveness of the entire extended auditor’s report instead of individual parts. Previous literature focuses mainly on particular paragraphs of the EAR such as the Key Audit Matters and the reaction of nonprofessional investors (Lennox et al.,2018; Christensen et al.,2014; Sirois et al.,2018), while this research will conduct the EAR as a whole. Furthermore, this research will explore whether the investment goal and risk preference of nonprofessional investors influence the extent of using the EAR. To our best knowledge, these particular relations are not subjected to previous research. This can be of favor for regulators and accounting firms since these parties are better able to assess who is using the EAR. For every relation to the EAR, we first test the particular relation with the annual report.

Practical relevance

In this study, the focus will be on one particular group of stakeholders namely: nonprofessional investors. The focus on this group is because there is expected that, in particular, this stakeholder group will benefit from the EAR. Nonprofessional investors rely mainly on publically available information, while other stakeholders often have more facilities and resources to guide their decisions (Asare and Wright, 2012). The amount of nonprofessional investors who invested in stock markets is approximately 1,4 million and growing in the Netherlands3. This research about the new information source is relevant to nonprofessional investors because, with a survey, we map the information needs of nonprofessional investors and possible poinst for improvement. Mapping the needs and points of improvement of the nonprofessional investors is relevant to the Dutch regulators and accounting firms as well since this research may provide more insight and improvement points on the information provision of the extended auditors report. It helps them to achieve the goal of the extended auditors report providing stakeholders with better and company-specific audit information (NBA, 2014).

3_{The number of investors is estimated by the staff of www.deaandeelhouder.nl, an information site which provides}

(9)

From the results, a significant negative effect was found on the relation between risk preference and reading the EAR which indicated that risk-averse investors are more likely to read the EAR. This might be caused by the fact that risk-averse investors want to be sure about their investment and therefore evaluate more information before deciding (Tseng and Yang, 2011).This study provides evidence for the moderating effect of education on the relation between risk preference and reading the EAR. This positive moderating effect can be explained by the fact that higher educated investors better understand complex information and do not evade primary information sources (Elliot et al., 2008; Chandra and Ravinder, 2011). With regards to the investment goal, no significant effect was found. According to the results, there is no difference in the use of the EAR when investing with retirement purposes or not investing with retirement purposes. This could be explained by the fact that the information which is preferred by investors with retirement purposes is provided in other information resources than the reports subjected to this study as well.

(10)

II. Theoretical background

2.1 Agency theory

In the past, the persons who owned the firm also controlled the firm. Because of the growing economy, companies wanted to attract more capital to finance their plans and expand their operations. Investors were needed to attract this capital, but in return, the investors wanted a share in the company (Salifu et al., 2015). The agency theory of Jensen and Mechling (1976) primarily relates to the relation between agents (the management) and principals (the shareholders or investors) of a stock listed firm. Principals own the firm but do not work at the firm. Therefore, principals appointed agents to run the company for them. The appointed agents should run the company in the best interest of the principal. There is a separation between ownership and control. This separation causes a gap of information provision since the principal does not know what is happening inside the firm. The firms’ agents do have more confidential information about the actual value of the firm compared to the principals, resulting in information asymmetry. Conflicting interests between agents and principals could result in agents that do not always act in the best interest of the principals.

Agency costs occur when the principal tries to limit the divergences, the costs of the audit are a parts of these costs (Jensen and Meckling, 1976). Agents can, for example, be more optimistic about the actual economic performance of the company than reality. The agents produce an annual report with financial statements to provide information about the current state and value of the company (Healy and Palepi, 2011). This is where the role of the auditors comes in. The primary task of the accounting profession is the verification of the financial statements (Chandler et al., 1993). However, as mentioned in the introduction, the value of the audit came under fire during the financial crisis. The ACCA (Association of Chartered Certified Accountants) in collaboration with the SIAS (Securities Investors Association Singapore) researched the value of the audit in the perspective of nonprofessional investors. In their research, “90% of the respondents felt that the external audit, in its current form and scope, brings value to them” and “80% of the respondents felt that the audited financial statements were important to them in making investment decisions” (ACCA, 2011). So, in the agency perspective, the audit is still an effective manner to reduce the agency conflicts since it enhances trust and reduces the information asymmetry between principal and agent. To communicate the result of the audit, auditors provide an auditor’s report. The auditor’s report is the medium to communicate the result of the audit to the public (Coram et al., 2011). The former auditor’s report was a single opinion but the new EAR adds company-specific information to the report. Regarding the agency theory, this should reduce the agency conflict even more. The inside information disclosed in the EAR should reduce the information asymmetry. This study builds on the agency theory since the EAR is a product of a demand for more inside information about the company. The EAR provides not only information about the work performed by the auditor but it is also a valuable addition at the total of information provided about the financial statements in the annual report (Brouwer, Eimers and Langedijk, 2016).

(11)

2.2 Legitimacy Theory

“Legitimacy is a generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions” (Suchman, 1995). To be able to operate, companies are looking to be legitimate in the relevant society. Companies havea social contract, which defines the boundaries to operate within society. The boundaries are determined by the norms and expectations of the relevant society (Deegan 2006). A company has to meet the expectations of society. Otherwise, the society will revoke the contract. When the expectations of society and the actions of the organizations are not in line with each other, a legitimacy gap will arise. Gray et al., (1995) are suggesting four strategies in which an organization can maintain the social contract and become or stay legitimate: 1) Educate and inform the society in which the organizations operate about the actual changes in the operations and behavior; 2) Change perceptions of the society in which the organization operates but not actually change the operations or behavior; 3) Manipulate perception by deflecting attention to other issues, organizations or areas; 4) Change the evaluation process by changing the expectations of the relevant society. Gray et al. (1995), suggest that “The first strategy is chosen in response to a recognition that the ‘legitimacy gap’ arose from an actual failure of performance of the organization” and “The second strategy is chosen as a response when the organization sees that the legitimacy gap has arisen through misperceptions on the part of the relevant public”. With regards to the EAR, it can be appointed that both strategy one and strategy two can be seen as a possible incentive for the regulators to maintain the social contract. On the one hand, writing the audit report is part of the job, and it can be argued that the gap arose from an actual failure. On the other hand, it can be argued that regulators mandated accounting firms to provide more information, but they had not to change the actual operations. The EAR is, among other things, a result of a misunderstanding of society in the work performed by the auditors (ICAEW, 2017). Whether it is an actual change of behavior or just changing perceptions of society, the goal of regulators is clear. By informing the society about the work performed during the audit and disclose the matters encountered, accounting firms try to reduce the legitimacy gap.

2.3 Expectation gap

The two theories discussed above both focus on the relation between companies and their stakeholders/relevant society, and how information can enhance this relation. The audit expectation gap especially focuses on the relation between the auditor and their relevant society. Liggio (1974) is the founder of the audit expectation gap and after the introduction, the expectation gap is a widely discussed subject. Different adjustments to the original definition came up. The original definition of the expectation gap given by Liggio (1974): “A factor of the levels of expected performance as envisioned both the independent accountant and by the user of the financial statement. The difference between these levels of expected performance is the expectation gap”. Porter (1993) divided this definition into two components: the reasonableness gap and the performance gap. The reasonableness gap is the gap between the duties reasonably expected of auditors and what duties society expects from auditors to do. The performance gap is the gap between the perceived performance of auditors by the society and the duties which can

(12)

reasonably be expected of auditors. What is the link between the EAR and the expectation gap? The auditor’s report may perhaps help with closing the gap. To close the expectation gap, Nair and Rittenberg (1987) came up with the suggestion to add additional information in the auditor’s report. This solution was supported by various studies over the following years (Kelly and Mohrweis, 1989; Miller and Strawser 1990; Gay and Schelluch, 1993). “In general, these studies provided evidence that an expanded audit report gives a fuller understanding of the scope, nature, and significance of the audit and influences the reader’s perceptions concerning the audit and the auditor’s role” (Koh and Woo, 1998). Approximately 26 years after the first suggestion to add more information to close the gap, regulators introduced the new standard. On the one hand, it can close the expectation gap since it reduces the misperceptions of society in some dimensions. On the other hand, it can evaluate the audit function because society disagrees with the current duty of accountants, which may lead to a larger gap (Leung and Chau, 2001). The EAR is widely discussed to close the gap, this study will determine to what extent non-professional investors use the report . If the use of the report is scarce, the solution to close the expectation gap with the EAR might be reconsidered.

(13)

III. Prior literature and hypotheses

3.1 The extended auditor’s report

The audit report is part of the communication of the economic information about the audited company. “It is intended to add credibility to the financial representations of management to the extent it is believed to be justified” (Pound, 1982). However, adding credibility is not enough to be informative in the eye of stakeholders. Various researchers concluded that the audit report was not more than just a stamp, a boilerplate, that gives a pass or fail opinion. The former auditor’s report was evaluated as largely uninformative because there was a lack of company-specific information (Geiger, 1993; Smieliauskas et al.,2008; Church et al., 2008; Gray et al., 2011; Mock et al., 2013). The suggestion to add more information in the audit report has resulted in the introduction of the EAR became mandatory in the Netherlands from the fiscal year 2014 (NBA, 2014). The newly introduced paragraphs (Key Audit Matters, Materiality and Audit Scope) should provide investors with more client-specific information about the audit.

Although the introduction of the EAR is quite recent, and society had to get used to the new standard, the literature regarding the effects of the EAR is growing rapidly (Gutierrez et al., 2016; Sirois et al., 2018; Reid et al 2018; Lennox et al, 2018; Smith, 2016; Christensen et al., 2014). From the literature, there is not a uniform conclusion about the extent to which the EAR is used. Results of an experimental study of Carver et al. (2017), who studied only the effect of the Key Audit Matters paragraph, concluded that the extended auditors report fails to add value for investors. Lennox et al. (2018) performed their study with regards to the effect of the new auditing report by using equity valuation models at 489 U.K. companies with available relevant data. The same study also suggest that “investors do not consider the additional risk disclosures to be incrementally informative.” These conclusions are in line with the research of Gutierrez et al. (2016). In a difference in differences research design, they compared two groups of companies with different auditor’s requirements. They gave three possible reasons for the suggestion that the EAR is not providing additional information to investors. “(1) an auditor’s disclosures may be preempted by other information; (2) the market may believe that an auditor adequately addresses the disclosed risks, previously known to the audit committee; or (3) the market may not understand the implications of the auditor’s disclosure” (Gutierrez et al., 2016). There are not only studies that provide arguments in favor of the opponents of the ‘added value’ of the EAR. Over the last few years, researchers also found evidence in favor of the supporters of the EAR. The experimental research of Christensen et al. (2014) examines the reactions of nonprofessional investors in which business graduate are used as a proxy for nonprofessional investors. Participants received the traditional auditor’s report and the traditional auditors report with Key Audit Matters paragraph included. Christensen et al. (2014) concluded that nonprofessional investors who received an auditor’s report with the Key Audit Matters paragraph included are more likely to exclude the particular company from the list of potential investments. Reid et al. (2015) found in their study in the UK a significant increase of abnormal trading volume and they suggest that the additional

(14)

information provided in the auditor’s report is found informative by investors. With eye-tracking software in the UK, Sirois et al. (2018) studied the effect of the extended auditors report on how investors navigate through the financial statements. A significant effect was measured. Investors gave more attention to areas identified as risky in the EAR. Smith (2016) examined whether the introduction of the EAR influences the communication value in the UK. With readability and tone measures, Smith (2016) concluded that after the introduction of the new standard, the communication value of the audit report is improved. Prior research mainly focuses on the UK and highlight primarily on the Key Audit Matters paragraph. This research, we focus on the Netherlands and will not highlight a particular paragraph, but it will take the extended auditors report as a whole.

3.2 Relevance for nonprofessional investors

Nonprofessional investors can consult multiple information sources before they make their investment decisions. Cascino et al. (2014) concluded that the most used sources by nonprofessional investors consist of: “the use of public media, advice by financial institutions, friends or family and financial statements.” However, the process of bounded rationality explains why investors are not able to determine all the available information sources. The cognitive decision-making capacity of investors is not entirely rational because individuals cannot process all available information (Lipman, 1995). Therefore, it is likely that investors reduce their information sources to an organized set of information that they prefer to use (Coval and Moskowitz, 1999). Lawrence (2013) investigated the relation between nonprofessional investors and the use of financial disclosure of firms. Nonprofessional investors prefer to invest in companies who provide a higher quality of financial information. “Specifically, individual investors shareholdings increase in firms with more concise, readable and transparent financial disclosures” (Lawrence, 2013). When the companies’ financial information meets the relevant requirements, it is more likely for nonprofessional investors to invest in a company. The EAR is not only useful for providing transparency of the work performed by the auditor but it can also add value to the financial disclosure. It better provides the information accessibility and it provides client-specific financial risk information (Smith, 2016). The EAR would ‘‘provide a roadmap to help users better navigate complex financial reports and focus them on matters likely to be important to their decision-making’’ (IAASB, 2011).

Particularly paragraphs are introduced in the EAR to provide insight in “areas that investors have indicated would be of particular interest to them, such as significant management estimates and judgments made in preparing the financial statements; areas of high financial statement and audit risk; unusual transactions; and other significant changes in the financial statements” as stated by M. Baumann, chief auditor at the PCAOB4_{. As introduced in the introduction, the VEB sends a}

letter to all big accounting firms every year with their expectations and points of attention for the accountants. Since 2013, the VEB highlights the EAR as a point of attention. In their letters, they

4_{Statement of M. Baumann, Chief Auditor the PCAOB. Assessed on October 2, 2018. via:}

(15)

are positive about the developments regarding the EAR, and they make some suggestions for adjustments (VEB, 2013-2018). From these letters, it can be assumed that nonprofessional investors in the Netherlands find the EAR relevant. Otherwise, they would not be this positive.

3.3 Investment goal

Nonprofessional investors can have various reasons for investing their money. For example, building a retirement, maintaining retirement, striving for maximum growth of capital or storing excess cash. These different goals are often coupled with different investment horizon. Two general investment horizons can be distinguished, namely: long and short term. Papaioannou et al. (2013) defined long term investors as “those who have the intention of holding an asset for multiple years and are not expected to liquidate their positions in the short term,” short term investments are the opposite. In general, long-term investments are less risky than short term investments (Gray, 2006) and it is a “more fundamental, research-oriented investment approach that assesses all risks to the business” (Marathon Club, 2007). Short-term investments are, among other things, exposed to business cycle fluctuations and timing is crucial since investors have to sell them just at the right time before the price drops down (Shen, 2005). Warren (2014) concluded that the information used by investors depends, among others, on the investment horizon. In his research, Warren (2014) sets out the different types of information used by long- and short-term investments. Investors who focus on short term investment are interested in information that might drive price changes. Information about news flows, earnings announcements, actions of other investors, market themes and share price evaluations are examples of information used by short term investors. They are less focusing on long-term value drivers. Investors who prefer long-term investments are more interested in the ability to be profitable and continue operations of companies. This kind of information is shown in business profitability, growth opportunities, and management quality. Investors can find this kind of information in the annual report of a company (Li, He and Xiao, 2018). Because short term investments have a higher chance to fail, they are less suitable for retirement purposes. Therefore, it can be assumed that investors who invest to build up or maintain their retirement, prefer long-term investments. As mentioned, short-term investors prefer information about the current state of the company which drives price changes. Long-term investors prefer information that provides information about the performance over time. The annual report of the company gives an insight into the performance over time, and therefore the following hypothesis is set:

Hypothesis 1: Investors’ who invest for retirement purposes are more likely to make use of the

annual report of the company in the Netherlands than investors who do not invest for retirement purposes.

Besides the pass or fail opinion, the EAR provides information about the financial performance and future plans of the company. Accountants have to judge the continuity of the company in the EAR (NBA, 2014). Because the EAR might provide information that suits even more within the information preferences of investors who invest for retirement purposes, the following hypotheses is formulated:

(16)

Hypothesis 2: Investors who invest for retirement purposes are more likely to make use of the

extended auditors report in the Netherlands than investors who do not invest for retirement purposes.

3.4 Risk Preference

“Risk preference refers to the attitude people hold toward risks, which is a key factor in studies on investors’ decision-making behavior” (Wen, He and Chen, 2014). The risk tolerance of a nonprofessional investor depends on many factors. Dorn and Huberman (2003) concluded that financial healthiness and experience of investors leads to a less risk tolerant investors’ profile. Grable (2017) defined risk tolerance with regards to financial decision making as “the maximum amount of uncertainty someone is willing to accept when making a financial decision.” According to the preferred risk habitat hypothesis of Dorn and Huberman (2010) the risk portfolio of the investor is linked to the volatility of the return of the investment. High volatility in returns would indicate a risky investment. When investors prefer a higher risk, they are more risk tolerant. The risk tolerance influences the decisions investors make with regards to their investments. As mentioned in the previous paragraph, the goal of investments varies from extra savings to a provision. These goals can indicate the risk that an investor is willing to take (Grable, 2017). Those who are more risk tolerance, are more aggressive in their investments. Risk-tolerant investors believe that they control the higher risk of the investment by a high frequency of short-term investments. This controls the risks of losing large amounts because they sell their assets before a large price drop.

In contrast to risk-tolerant investors, risk-averse investors prefer more certainty in their investments. Satti et al. (2013) suggest that investors who want to reduce the uncertainty of the investment, have to spend more time in their information evaluation process. This is in line with research of Tseng and Yang (2011), their findings show that “individuals with more risk aversion tend to seek more information.” Nowadays, it is for individual investor easier to gain more information due to the digital era. The need for more information when investors want to reduce the uncertainty is the fundament of the third hypothesis. The annual report is a detailed source of information, which shows the financial and organizational performance of the company. This research assumes that investors who read the annual report, are better able to generate a better overall picture of the company. Thereby, the risk disclosures in the annual report could risk-averse investors indicate whether it fits within their risk profile. The third hypothesis is formulated as followed:

Hypothesis 3: The extent to which an investor is risk averse increases the likelihood of using the

annual report of the company in the investor’s decision-making process in the Netherlands

The EAR provides the user with more information about the audit of the financial statements. As mentioned, risk-averse investors tend to seek more information. Besides that, the EAR can be helpful as guidance for cutting through the complexity of the financial statements (IAASB, 2011); the report provides more information about the company. The Key Audit Matter paragraph, for

(17)

instance, might provide investors information about risky high debt levels or another risk the auditor encountered during the audit (Porumb et al., 2018) This is valuable information for nonprofessional investors who are seeking to reduce the uncertainty in their investment. This assumption is the basis of the fourth hypothesis:

Hypothesis 4: The extent to which an investor is risk-averse increases the likelihood of using the

extended auditors report in the investors’ decision-making process in the Netherlands.

3.5 Education

The traditional finance theory suggests that individuals are fully rational in their investment decision process and they consider all available market information (Suryawanshi and Jumle, 2016). However, several studies concluded that investors are irrational and overconfident (Dorn et al., 2003; Murthy and Joshi, 2012; Grover and Singh, 2015; Bakar and Yhi, 2016). As mentioned, risk-averse investors seek more information. Natural factors can cause the risk-averse behavior of investors such as gender, race or it is in more or less strongly in someone’s genes (Byrnes et al., 1999; Manfredi et al., 1992; Cesarini et al., 2009). Risk aversion can also be influenced by nurture determinants, for example education (Dohmen et al., 2005; Hrysko et al., 2011). Education would make investors more comfortable with risks. Educated individuals better understand how to deal with risks. This is one of the things they learned during their study (Jung, 2014). Besides that, educated individuals are able to better evaluate information. Well-educated individuals can process information more easily and have learned to be more rational (Moshman, 1990). Poorly educated investors have more difficulties with understanding the financial statements, and so the accessory extended auditors report, these investors will more avoid this kind of information (Chandra and Ravinder, 2011).

Elliot et al. (2008) suggest in their study that education has a positive relation regarding the usage of primary information sources (information directly disclosed by the company). Taking this information together, it can be assumed that education is a moderator with a positive influence on the relation between risk-averse investors and the extent of reading the EAR. This is expected because well-educated investors are less avoiding the information in the financial statements and the accessory EAR. Therefore, the final two hypotheses are as following:

Hypothesis 5: Education positively moderates the relationship between the risk preference of

nonprofessional investors and the extent to which they use the annual report of a company in their decision-making process in the Netherlands.

Hypothesis 6: Education positively moderates the relationship between the risk preference of

nonprofessional investors and the extent to which they use the extended auditors report in the Netherlands.

(18)

IV. Methodology

4.1 Design

Instead of looking at company data (Lennox et al., 2018), performing an experiment (Christensen et al., 2014; Smith, 2016) or using data from abnormal trading volumes (Reid et al., 2018), a survey is used to gather our data. Together with my fellow researchers, who are studying the effect of the EAR as well, we developed a questionnaire with relevant questions for our studies. In his paper, Kreamer (1991) presented the three main characteristics for performing survey research. First of all, when describing particular aspects of a population quantitatively. Secondly, when the required data have to be collected from the opinions of people. Thirdly, when the research findings from the sample have to be generalizable over the whole population. Besides the fact that those three characteristics fit within our research design, we preferred the survey instrument because our research has an explorative character. We had to generate data by ourselves since there is not a lot of relevant data available which could answer our hypotheses. Since we study particular preferences and characteristics, survey research fits the best within our research. Figure 1 provides an overview of the different tests we used for our hypotheses.

4.2 Population and Sample

Our research is about the value of the extended auditors report from a non-professional investors perspective. Therefore, our population is aimed at nonprofessional investors. Nonprofessional investors are a heterogeneous group of individuals who might have completely different purposes and characteristics with regards to investing (Elliot, Hodge, Kennedy and Pronk., 2007). Elliot et al. (2007) concluded that M.B.A. students are a good proxy for nonprofessional investors but the information acquiring, and integration process might be different, especially for tasks with a relatively high integrative complexity. Therefore, we have chosen to consult actual nonprofessional investors for our research to provide more reliable findings. We did this by contacting the VEB (the association of nonprofessional investors in the Netherlands) via the guiding professor of this research, who is a member of the supervisory board of the VEB. The VEB protects its member's interest by providing support and independent information about investment options. “Besides being the main source of independent information for the Dutch investor, the VEB is widely known for its collective redress – lawsuits typically brought to obtain compensation for groups of aggrieved shareholders.”5 The VEB is representing 41.841 members according to their annual report in 2017 (Annual Report VEB, 2017). Becoming a member of the VEB includes several benefits such as: being eligible for the benefits of the lawsuits of the VEB, receiving a magazine, an online newsletter and being allowed to participate in various investment-events organized by the VEB. Therefore, we assume that every individual who is seriously interested in investing is a member of the VEB. It is hard to determine which part of the VEB members is still an active investor and therefore we keep all the members of the VEB as our population. A direct consequence of holding a population this large, our response rate was quite

(19)

low. In total, 132 nonprofessional investors participated in our survey. After a completeness check, 118 investors where left over. With the formula of Smith (2013), we can determine the minimum required responses without exactly knowing the size of our population to run statistical tests. According to Smith (2013), we can conclude that a sample of 118 is enough to test our hypotheses6_{. To boost the sample size, we used a method described by Sheehan (2001): we have}

sent reminders to fill in the survey to the member of the VEB twice. Besides that, we have tried to keep the survey as short as possible in order to prevent responders from stopping halfway.

4.3 Survey Instrument and Execution

To reach our population, we gave all VEB members the possibility to contribute to our survey research. In the online weekly newsletter, we provided a button for our survey. We asked the members of the VEB if they could fill in the survey and add a contribution to science to trigger them. Besides the scientific contribution the participants make, we explained to them that it could work in their benefit since they help to improve the information distribution of organizations. To distribute our survey, we used Qualtrics which is an online platform providing various applications for designing, distributing and analyzing surveys. Everyone could fill in the survey since the link was provided on the website of the VEB as well. We assume that a self-selection process ensured that only nonprofessional investor filled in our survey. We clearly formulated the purpose and the aimed population of the survey so that it would disattract individuals who would not fit within our population. According to Levy and Lemeshow (1999), issuing a pilot survey is strongly recommended. We issued a pilot survey for three fellow students before we executed the survey on a large scale to check the efficiency and verify the questionnaire. We did not have space in our survey to add additional questions for reliability purposes. Since we had to include three single studies and some additional questions for further research we could only ask the highly relevant questions. Additional reliability questions would make the survey too long, resulting in a lower response rate. We assumed the questions were sufficient straight-forward to ensure reliable responses.

4.4 Study variables and tests

Reading the annual report of the company (READING AR): Based on survey research of Khan,

Tan, and Chong (2017) we drew up a list of the most common information sources. We added information sources that were relevant for our research, in special ‘the annual report of the company.’ We asked the respondents on a five-point scale, from never to very often, in which extent they used the particular information source (please refer to Q15, Appendix 1). We measured this variable as a scale variable. We transformed the values into z-scores to create a better understanding of the sample and to be able to compare these variables with other variables. An

6_{Smith (2013) introduced an equation for an unknown population size or a large population size. Since we do not}

exactly know what the population size is, but we do know that the population is large, we used the equation of Smith (2013): Necessary Sample Size = (Z-score)2 * StdDev*(1-StdDev) / (margin of error)2. We took a confidence level of 95% and a margin of error of 9%.

(20)

Ordinary Least Squares regression is used to determine the effect of the independent variables on READING AR.

Reading the extended auditors report (READING EAR.): To measure the EAR, we used a

follow-up question method. First, we asked which information sources the respondents are using for their decision-making process (please refer to Q 15, Appendix 1), and as follow-up question: ‘If you would read the annual report, which specific parts would you read?’ (please refer to Q16, Appendix 1). One of the options is the EAR. In this way, we can conclude whether or not the EAR is being read. The follow-up questions ensured that only respondents who read the annual report can read the EAR. This dependent variable is measured as a categorical variable. The respondent scores one point, only when the respondent used the annual report as an information source (Q15) and does read the extended auditors report in the follow-up question (Q16). They scored zero points otherwise. Since this is a discrete variable, we have used a logistic regression to determine the effect of the independent variables on READING EAR.

Investment goal (INV. GOAL): To measure whether or not the respondents invest for retirement

purposes, we asked them to select one or more investment goals (please refer to Q10, Appendix 1). We provided the four most common investment goals to the best of our knowledge. Although those four goals covered almost the entire sample, we added the option “otherwise, namely.” A few respondents answered ‘‘just for fun’’ but with the four options we have covered the most significant part of the population (90%). From the four options, there are two retirement-related possibilities, namely: building retirement and maintenance of retirement. When the respondent selected one of those two retirement-goal options, they scored one point, and otherwise zero. Therefore, this independent variable is measured as a categorical variable.

Risk Preference (RISK PR.): To measure the risk preference of investors, we asked the respondents

to rate themselves on a scale from 1 to 9 if they are willing to take more risk in return for a higher return (please refer to Q9, Appendix 1). This is in line with the study of Dohmen et al. (2005) who concluded that this general question gives a good indication of a person’s risk profile. We transformed the risk scores into z-scores to better understand the values and to be able to compare this variable to other variables.

Education (EDU): Education is measured straightforward, with the question: “what is your highest

level of education you have completed” (please refer to Q4, Appendix 1). This interaction variable is measured as categorical variable. Respondents who completed their study at a research university or a university of applied science (in Dutch: hoger beroepsonderwijs) are labeled as ‘higher educated,’ and ‘lower educated otherwise.

(21)

Figure 1: Overview of the different tests we used to analyze our hypotheses.

4.5 Control variables

We added control variables to ensure that we only measure the effect of the independent variables. Moreover, we included variables that prior research suggests having an influence on the usage of information by non-professional investors. We control for AGE since this can influence the behavior regarding investments (Elliot, 2006). The ages of respondents are not normally distributed, so we categorized this variable into three groups to make them normally distributed and ready to perform tests with. The first group included the ages from >20 to 60, the second group included ages from 60 to 70 and the last groups included ages older than 70. We included the dummy variable GENDER (male=1, female =0) to control for gender differences in investment behavior (Barber and Odean, 2001). Furthermore, we control for experience

(EXPERIENCE) since Elliot (2008) provided evidence that an investors’ experience might influence the use of complex information. We transformed this variable from five groups to three groups since it was not normally distributed. The first groups included respondents with

investing experience from 0 to 20 years; the second group included 20 to 30 years and the last group included 30 years or more. An overview of all variables is provided in table 1.

(22)

Variable Description Independent variable

READING AR Scale variable (from 1 to 5) to which extent the annual report is being read. READING EAR Dummy variable whether the extended auditor's report is being read yes or no. Dependent variable

RISK PREF

Scale variable (from 1 to 9) to which extent the respondent is willing to take risk.

INV. GOAL Dummy variable whether or not the respondent invests for retirement purposes. Moderator

EDUCATION Dummy variable whether or not the respondent is higher educated. Control variables

AGE

Categorical variable in which the age of the respondent divided into three age groups.

GENDER Dummy variable for the gender of the respondent, male or female.

EXPERIENCE Categorical variable divided into three groups of years’ experience in investing.

Table 1: Overview and a short description of all variables

4.5 Additional analyses

To generate extra context about this research, we analyzed the opinions of the nonprofessional investors towards the EAR as well. In the survey, we asked the respondents which paragraphs they would read of the EAR (please refer to Q17). In here, we analyze which paragraphs the readers of the EAR do read and which paragraphs they do not. We also analyzed the difference between the groups we used for the regressions. We divided the respondents in three risk preference groups: group 1 has a risk preference of 1 to 3, groups 2 has a risk preference from 4 to 6 and group 3 has a risk preference from 7 to 9. Also, we divided the respondents under a group who invest with retirement purposes and a group who does not invest with retirement purposes.

Second additional analyses are performed regarding the question what kind of information should be added to the EAR to be of value for them (please refer to Q18). We have executed the same analyses on this question compared to the question which paragraphs the respondents would read.

(23)

V. Results

5.1 Descriptive statistics

In this chapter, we will discuss the results concerning the hypotheses formulated in this research. From the 133 respondents who started the survey, 106 finalized it. Since the survey included questions for other studies as well, not all questions had to be answered for this particular study. In the end, 118 respondents (Table 2) filled in the questions required for this study. The sample consists of 113 males (MEAN = 0,96) and 5 females. The average age was between 50 and 60 years (MEAN = 4,68) old, but most of the respondents were 60 to 70 years old (MODE = 6). The average experience of the respondents is between 20 and 30 year (MEAN = 3,76). The age and experience are in line with the research of Taylor and Riley (2017) who aimed their research on nonprofessional investors as well. Most of the respondents are highly educated (MEAN = 0,74). With regards to investment decisions, the respondents indicated that in general, they read the annual report rarely to sometimes (MEAN = 2,66). 39% (MEAN= 0.39) of the respondents read the EAR. The respondents had a risk preference with a mean of 6,25 and a mode of 7 on a scale from 1 to 9. 29% of the respondents are investing for retirement purposes (MEAN = 0,29).

N Mean Mode Min Max Std. Deviation

READING AR 118 2,66 3 1 5 1,24 READING EAR 118 0,39 0 0 2 0,49 INV. GOAL 118 0,29 0 0 2 0,46 RISK PREF. 118 6,25 7 1 9 0,73 EDUCATION 118 0,74 1 0 1 0,44 AGE (3 groups) 118 2,11 3 1 3 0,83 AGE (6 groups) 118 4,86 6 1 6 1,3 GENDER 118 0,96 1 0 1 0,202 EXPERIENCE (3 groups) 118 1,97 2 1 3 0,82 EXPERIENCE (5 groups) 118 3,76 4 1 5 1,17

Table 2: Descriptive statistics of dependent, independent and control variables

Table 3 provides the correlation matrix. We tested the correlation with both the Pearson correlation and the Spearman correlation test. As we can see, reading the annual report and reading the EAR are positively significant correlated (r = 0,471). This is quite logical since the EAR is provided within the annual report. Experience is positively significantly correlated with reading the EAR (r

= 0,182), in which we can conclude that more experienced investors are more likely to read the

extended auditors report. Furthermore, experience is positively correlated to risk preference (r = 0,191), age (r = 0,191) and gender (r = 0,199) as well. More experienced investors take higher risks according to the results. The correlation between experience and age was expected since more experienced investors obviously are at an older age. Gender is significant, which would indicate that more experienced investors are more likely to be male but since only 5 females filled in the survey, we cannot draw a proper conclusion with regards to gender.

(24)

C o rr e la ti o n s m at ri x iab le 1 2 3 4 5 6 7 8 D IN G AR - - D IN G EAR ,4 8 2 ** (,4 7 1 ** ) - - V. GOAL ,2 2 (,0 2 4 ) -,0 1 (-0 ,0 1 ) - - K P REF. -,0 3 6 (-,0 1 7 ) -,0 8 (-,0 5 4 ) -,1 4 -( ,1 6 7 ) - - D UCA TION ,0 0 9 (,0 0 1 ) ,0 4 3 (,0 4 3 ) ,0 8 2 (,0 8 2 ) ,0 9 (,1 0 7 ) - - E -,0 0 2 (,0 7 8 ) ,0 8 2 (,0 8 3 ) -,0 3 9 (-,0 3 9 ) -,2 3 1 (-,2 2 9 ) -,1 5 3 (-,1 52 ) - - D ER ,1 5 (,1 4 9 ) ,1 6 8 (,1 6 8 ) -,1 4 5 (-,1 4 5 ) -,0 8 1 (,1 0 1 ) ,0 66 (,0 66 ) -,0 73 (-,0 66 ) - - P ER IEN CE ,0 6 9 (,0 7 8 ) ,1 8 3 * (,1 8 2 * ) -,1 1 2 (-,1 1 1 ) ,1 7 (,1 91 * ) ,1 65 (,1 63 ) ,1 94 * (,1 91 * ) ,1 98 * (,1 99 * ) - - o rre la tio n is sign ifi can t at t h e 0 ,0 1 le vel. rr elati o n is sig n ifi can t a t the 0 ,0 5 lev el. 3: C o rr e lation m atr ix o f d e p e n d en t, i n d e p e n d e n t an d c o n tr o l var iab le s. F ir st val u e ar ises fr o m t h e Pe ar son c o rr e lation t e st, t h e ( seco n d ) val u e ar ises fr o m t h e Sp e ar m an lation t e st.

(25)

5.2 Linear Regressions

Table 4 presents the results of the first five models of this study. These models are tested by linear regression analyses and give the results for hypotheses 1, 3 and 5. The other hypotheses are tested by logistic regression in paragraph 3.

Model 1 includes only the control variables. Model 2 includes the control variables with the first independent variable, INV. GOAL. The third model includes the control variables with the first and the second independent variable RISK PREF and INV. GOAL. The fourth model includes the control variables, the two independent variables, and EDUCATION to measure the main effect first. The last, fifth model includes the control variables the two independent variables, the variable EDUCATION and the interaction variable EDUCATION x RISK PREF.

Model 1: The first model only includes the control variables. We made the assumptions for the

control variables derived relevant literature regards information sources. However, there is no significant effect on the control variables AGE (Middle vs. Low 𝛽 = -.0,251, 𝜌 > 0.1 and High vs. Low 𝛽 = -.0,019, 𝜌 > 0.1) GENDER (𝛽 = .624, 𝜌 > 0.1) and EXPERIENCE (Middle vs. Low 𝛽 = -0.093, 𝜌 > 0.1 and High vs. Low 0,126) on READING AR. This means that there is no relation between the control variables and READING AR.

Model 2: H1 predicted that investors who invest for retirement purposes would make more use of the annual report of the company. When investing for retirement purposes, the value of INV. GOAL is 1, and 0 otherwise. This hypothesis is analyzed in model 2. There is no significant relation between these two variables, therefore, based on the results, hypothesis 1 is not supported (𝛽 = .114, 𝜌 > 0.1). This means that there is no effect of INV. GOAL on READING AR.

Model 3: H3 predicted that when an investor is more risk-averse, the investor would make more

use of the annual report. This hypothesis is tested by model 3. There is no significant effect of RISK PREF. on READING AR. Based on the results, hypothesis 3 is not supported. (𝛽 = -0.061, 𝜌 > 0.1). This means that RISK PREF. has no effect on READING AR.

Model 4: Before we measured the moderation effect of the interaction variable, we added

EDUCATION in model 4 to measure this variable as main effect on READING AR. EDUCATION has no significant effect on the dependent variable READING AR (𝛽 = -.024, 𝜌 > 0.1). Meaning that there is no direct effect of EDUCATION on READING AR.

Model 5: The fifth model measures the moderation effect of EDUCATION on the relation

between RISK PREF. and READING AR. Model 5 tests hypothesis 5 of our study. We predicted that EDUCATION positively moderates the relationship between RISK PREF. and READING AR, measured with the interaction variable EDUCATIONxRISK PREF. The interaction variable is positively significant with a 10% confidence interval (𝛽 = .386, 𝜌 < 0.1), which is in line with our hypothesis. Education has a positive effect on the relation between RISK PREF. and READING AR. Hypothesis 5 is accepted, but with a R2 of 0,078 and a 10% confidence interval, the effect is small.

(26)

Since we used linear regressions to test the hypotheses, we tested some assumptions as well. We tested for multicollinearity, autocorrelation and we checked for outliers. 1) Multicollinearity. A perfect linear relationship between independent variables must be prevented. Therefore, the independent variables should not correlate to high. We used VIF-scores to measure

multicollinearity since these scores indicate the linear relationship. There are no set rules about the VIF-value, but Meyers (1990) suggest that there is multicollinearity when the VIF-value equals 10. The highest VIF-value is reported in table 3. The highest VIF-value is 4,021 which indicate that there is no risk for multicollinearity. 2) Autocorrelation. We tested for

autocorrelation with the Durbin-Watson test. The Durbin-Watson test measures the presence of autocorrelation in the errors of the regression model. A value lower than 1 or higher than 3 is a reason to assume that the errors are correlated (Durbin and Watson, 1951). In our linear

regression model, the Durbin Watson score equals 2,282. Therefore, there is no autocorrelation in our regression model. 3)aCheck for outliers. We have checked the outliers whether these data

points had a significant influence on our regression results. We did this by looking to the Cook’s distance, which measures the effect on the results when we delete a certain data point. The Cook’s distance should not exceed 1. In our model the maximum Cook’s distance equals 0,176 (M = 0.078, SD = 0.019). This indicates that there are no data-points that affect our regression model.

Linear regression for Reading the Annual Report.

Model 1 Model 2 Model 3 Model 4 Model 5

Constant -0,523 -0,593 -0,576 -0,463 -0,447

Control Variables

AGE: Middle vs. Low -0,251 -0,352 -0,279 -0,284 -0,269

AGE: High vs. Low -0,019 -0,014 -0,055 -0,064 -0,127

EXPERIENCE: Middle vs Low -0,093 -0,092 -0,075 -0,074 -0,066

EXPERIENCE: High vs Low 0,126 0,136 0,167 0,179 0,208

GENDER 0,624 0,660 0,654 0,523 0,525

Independent Variable

INV. GOAL - 0,108 0,091 0,097 0,114

RISK PREF. - - -0,061 -0,061 -0,350*

EDUCATION - - - -0,05 -0,024

EDUCATION x RISK PREF. - - - - 0,395*

R2 0,043 0,045 0,049 0,049 0,078

ΔR2 0,043 0,002 0,003 -0,022 0,000

F-value 0,9888 0,863 0,789 0,690 0,997

N 118 118 118 118 118

Note: *p<0.10, **p<0.05, ***p<0.01.

Table 3: Linear regression results for hypotheses 1, 3 and 5. Unstandardized coefficients of the linear regression of READING AR.

(27)

5.3 Logistic Regressions

The dependent variable which measures whether the EAR is being read is READING EAR. Our second dependent variable is a discrete variable (0,1), and therefore, we used logistic regression in order to formulate an answer to our remaining hypotheses (hypothesis 2, 4 and 6). Hypothesis 2 predicted that nonprofessional investors who invested for retirement purposes, measured with INV. GOAL, are more likely to read the EAR. The results of hypothesis 2 (see: Table 4) show that there is no significant effect between INV. GOAL and READING EAR (β=0,063, p=0,883, eβ=1,066). With regards to the control variables, the results show that both age dummies are non-significant (Middle vs. Low: β=0,054, p=0,918, eβ=1,055 and High vs. Low: β=0,29, p=0,551, eβ=1,336) and so, do not affect whether or not the respondent reads the EAR. The control variable EXPERIENCE is non-significant for Middle vs. Low (β=0,003, p=0,995, eβ=1,003) but is significant for High vs. Low but only with a 90% bootstrap confidence interval (β=0,880 p=0,075, eβ=2,411). So nonprofessional investors with high experience, more than 30 years, in investing are 2,41 times more likely to read the EAR. The first logistic regression has a Nagelkerke R2 of 0,062, which indicated that the variables included in the model can explain approximately 6.2% of the variation in READING EAR. The Hosmer and Lemeshow test is non-significant (p=0,270) which indicates an acceptable model fit since there are no significant differences between the actual and predicted values.

Predictor β SE β Wald’s χ2 df p eβ (odds ratio)

Constant -0,993 0,562 2,76 1 0,097 0,393

Control variable

AGE: Middle vs. Low 0,054 0,519 0,11 1 0,918 1,055

AGE: High vs. Low 0,29 0,486 0,356 1 0,551 1,336

EXPERIENCE: Middle vs. Low 0,003 0,486 0,000 1 0,995 1,003 EXPERIENCE: High vs. Low 0,880 0,494 3,169 1 0,075* 2,411

Independent variable

INV. GOAL 0,063 0,432 0,022 1 0,883 1,066

Goodness-of-fit test

Hosmer and Lemeshow n/a n/a 9,931 8 0,270 n/a Note: *p<0.10, **p<0.05, ***p<0.01.

Table 4: Logistic Regression results for hypothesis 2. Coefficients of the model predicting whether a respondent reads the EAR. N=118 and Nagelkerke R2 = 0,062.

Adding additional information to the auditor’s report: The more, the better? Survey research among nonprofessional investors.