Effect of different economic conditions on audit and non-audit fees : evidence from the United Kingdom

(1)

Amsterdam Business School

MSc Thesis

Effect of different economic conditions on audit

and non-audit fees: Evidence from the United

Kingdom

Student

Name: Sema Ozcan

Student number: 10254552

Supervisor: dhr. dr. A. (Alexandros) Sikalidis

MSc Accountancy & Control, specialization Accountancy, 2013/2014

Faculty of Economics and Business, University of Amsterdam

(2)

2 Abstract:

During the recent financial crisis auditors receive extraordinary high fees from their clients. In this study I will use determinants from previous studies to investigate whether economic conditions have an impact on the amount of audit and non-audit fees. Understanding key drivers that determines the amount of audit and non-audit fees will help companies to increase or decrease the payment for fees. Also the setting of this study is distinctive from prior literature. Prior literature focuses on one year, while this study takes three time periods namely: favorable (2006-2007), economic downturn (2008-2009) and post-crisis (2010-2011). The objectives of this study is: (i) to investigate whether there is positive relation between audit and non-audit fees in the United Kingdom; (ii) to examine whether different economic conditions may influence this relationship; (iii) to discover whether economic conditions affects the ratio of audit and non-audit fees. These studies are based on publicly traded firms operating in the UK (financial institutions and utilities excluded) and are selected from the FTSE All Share Index from the period 2006-2011. The results indicate that there is a significant positive relationship between audit and non-audit fees. Second, I found a significant change in the relationship between audit and non-audit fees when economic conditions are changing. This is also the case for the change of ratio of audit and non-audit fees. The results indicate that moving from a period that the economy is doing well to a less favorable period that there is a significant change of the ratio of audit fees as well non-audit fees.

Key words: Audit fees, non-audit fees, economic conditions, financial crisis, ratio audit fees, ratio non-audit fees, United Kingdom.

(3)

3

1. Introduction

1.1. Background

The financial crisis of 2008 is considered to be the worst debacle that is experienced since the Great Depression. The economic meltdown started in the US as millions of people lost their job and trillions of dollars in market value disappeared. The major and most direct cause for the economic meltdown is the US housing market. Political intervention made home ownership more comfortable and banks were given excessive credits to home buyers. The bubble of given credits, that could not be pay back anymore, become bigger and bigger and in 2008 eventually this bubble burst. This resulted in a dramatic increase of interest rate which, in turn, had a great impact on global stock markets. The collapse put the world in a regression.

Figure 1: Global GDP Growth (Percent, quarter over quarter, annualized). Source: IMF, 2012

Figure 1 shows the major drop in the economic activity during 2008-2009 for the whole world and different kind of economies. The significant decline in the economic activity led to credit tightening and to a substantial decrease of international trade (Chor & Manova, 2012). As productions declined many companies had no choice than to, as much as possible, reduce any expenses. Many companies went bankrupt and millions of people lost their job.

The United Kingdom is also affected by the global crisis. The United Kingdom had known a period of unbroken growth since 1992, but in 2008 this dramatically changed. The credit tightening and job fears have resulted in weakened confidence of the consumers. Consumers decide to pay off debt and save. This led to a negative GDP Growth in 2008 in the UK. 1

The banking sector is blamed for the financial crisis, but also the audit profession is questioned. Banks, hedge funds and insurance are estimated to have lost around US$2.8 trillion (Bank of England, 2008) while the auditor sector earned millions from their services. The main concerns and critique on

1

(5)

5 the audit profession are the payments of audit fees. During the economic downturn auditors received extremely high fees from their clients. As the fees increases, the auditor can become less independent because of the economic bond these fees create between the client and auditor (Holland & Lane, 2012). The study of Sikka (2009) have showed that a great number of companies have failed in a short time after receiving unqualified audit reports. An unqualified audit report is issued when the auditors believes that the statements are free from material misstatements. If the auditor is convinced that there are certain misstatements, this will result in a qualified opinion. Many financial companies in the US, the UK, France, Germany, Switzerland, Iceland and the Netherlands received unqualified audit opinion prior to their statement of financial difficulties (Friedman & Friedman, 2009; Sikka, 2009).

Audit fees are defined by the Securities Exchange Commission (SEC, File No. PCAOB-2003-03) as "professional services rendered for the audit of an issuer's annual financial statements and (if applicable) for the reviews of an issuer's financial statements included in the issuer's quarterly reports”. Non-audit fees are defined as "all other services other than audit services, other accounting services, and tax services" (SEC, 2003). Different studies have tried to understand what determines high fees. The study of Simunic (1980) argues that different factors can be associated with audit costs and the results of his study indicate that audit price is associated with client size, complexity and risk. Other studies confirm that the most powerful impact on audit fees is client size (Taylor & Baker, 1981; Simon &Francis, 1988: Butterworth & Houghton, 1995). Complexity is also an important determinant of audit fees; the more complex the audit, the more time and effort is required. The determinant client risk is associated with client profitability. Another study of Simunic (1984) suggests in the first place that audit and non-audit are negatively correlated. This is based on the idea that joint provision of both audit and non-audit services can be more efficient since rendering once service the auditor obtains knowledge. However, the results revealed a positive relationship between two services, but according to Simunic (1984) this is due to ‘beneficial knowledge spillover between services (Simunic, 1984, p. 669). The measurement of factors influencing the audit fees and the relation between audit and non-audit fees will be extensively discussed in the literature review.

1.2. Research question

This study investigates the relationship between audit and non-audit fees in the UK and whether different economic conditions affect the relationship between audit and non-audit fees. Also the ratio of audit and non-audit fees will be investigated. The study is based on three periods representing different economic conditions namely: favorable (2006-2007), economic downturn (2008-2009) and post-crisis period (2010-2011).

(6)

6 The most common audit terms in this study used are:

 Audit service – evaluation of the relevance and the reliability of the financial statement of an organization.

 Non-audit service - include consultancies about tax, system, finance, investments, management as well as conducting of international business (Firth, 1997, p.513).

 Positive relationship between audit and non-audit fees (positive correlation between audit and non-audit fees) means that clients, who purchase more non-audit service, pay more for audit service.

 Ratio of non-audit fees – non-audit fees to total fees (DeFond, et.al., 2002)

 Ratio of audit fees – audit fees to total fees (DeFond, et.al., 2002)

 Big Four - the largest audit firms PricewaterhouseCoopers (PwC), Ernst & Young (EY), Deloitte and KPMG

The results suggest that there is a positive relationship between audit fees and non–audit fees and that this relationship becomes stronger in times of economic crisis. The trend of the ratio of audit fees is increasing while the ratio of non-audit fees is decreasing.

1.3. Contribution of thesis

This paper contributes to existing knowledge in number of ways. First, most of previous studies are conducted in the USA. This study provides evidence concerning the association between audit and non-audit fees in a different audit environment. Understanding of key audit fee drivers in UK setting can lead to efforts of companies to increase the payment for audit fees. Second, this study is conducted over three different time periods: favorable 2006-2007, economic downturn 2008-2009 and post-crisis 2010-2011. Also the total time period is taken into consideration 2006-2011. Previous studies only focused on one year. This may lead to inconsistent results. Also this study compares the correlation analysis of three different time periods to see whether economic conditions may affect the relationship between audit and non-audit fees. Third, by investigating the change in audit fee ratios during different economic conditions, we can see how economic conditions may influence this ratio in the future. As last, I have investigated the sensitivity of different industries to see which industry is the most sensitive to changing economic conditions.

1.4. Structure

The structure of the paper is as follows: Section 2 provides the relevant literature and hypothesis development. Section 3 describes the methodology and sample selection. In section 4 the results are described and section 5 provides the conclusion.

(7)

7

2. Literature review

2.1. Relationship audit and non-audit fees

Audit fees are defined by the Securities Exchange Commission (SEC, File No. PCAOB-2003-03) as "professional services rendered for the audit of an issuer's annual financial statements and (if applicable) for the reviews of an issuer's financial statements included in the issuer's quarterly reports” and non-audit fees as "all other services other than audit services, other accounting services, and tax services" (SEC, 2003). The last decade there is a discussion going on whether the auditor should also provide non-audit services to a client. It is argued that that supplying both services can create a conflict of interest and can be viewed as a threat to auditor independence. According to the Metcalf Committee Staff Study

(

1976, pp. 50-52) there can be conflict of interest, especially when the audit firm installs a management information system and subsequently audits the reliability and accuracy of its own work. However, contrary to this argument the Cohen Commission failed to find any significant relationship between the provision of management services and substandard audits (Simunic, 1984, p. 679). In the UK it is allowed to provide both audit and non-audit services as long as the independence of the auditor will not be affected.

The study of Simunic (1984) suggested that audit and non-audit fees services are negatively correlated. The reason for this is that joint provision of both audit and non-audit services can be more efficient. By providing the two services, the auditor obtains knowledge about the client’s business, which can improve the quality of both services. The synergy effect of providing both audit as well non-audit services is also called the “knowledge spillover” by Simunic. Contrary to the expectations of Simunic, the results provide evidence that there is positive relationship. It appeared that audit fees for client who purchase non-audit fees services from the auditor turn out to be higher than for clients who do not do this. In this study I investigate whether there is a positive relationship between audit and non-audit fees. That means that there is a positive relationship between non-audit and non-non-audit fees when client who purchase more non-audit services, pay more for the audit service.

The afore mentioned study of Simunic (1984) is a fundamental study concerning audit and non-audit fees. Based on this study, other studies tested also the association between audit and non-audit fees and try to provide explanations for the association between audit and non-audit fees. Simunic (1984) examines whether the cost of producing two services are interdependent. He suggests that producing two services will results in “knowledge externalities or spillovers” and reduce the total cost for the client. According to Simunic (1984) the form of spillover can be complex. “specifically when knowledge may flow from auditing to non-auditing services, or non-auditing to auditing, or in both directions; the fixed cost, variable cost, or both cost of the services can be affected; the knowledge spillover may be client-specific of general”( Simunic, 1984, p. 681). In his study he compared Big

(8)

8 Eight firms where only audit is purchased with Big Eight firm where fees paid for both services. The results showed that the purchase of both audit as well non-audit services from the same supplier were significantly higher for client than client that only purchase one service. This can be explained either by knowledge spillover or by systematic differences between audit purchasers and non-purchasers. Simunic (1984) highlights the importance of competition on the market. In high competitive market the profit of providing both services can be passed on to the client, while in less competitive markets the profit can be retained by the auditor.

Followed this research other studies have tried to explain the association between audit and non-audit fees. Palmrose (1986) investigate US public and closely-held companies, which were audited by Big Eight. Consistent with Simunic (1984), Palmrose also found a positive relation between audit fees and non-audit fees. Palmrose (1986) suggests that the correlation can be explained by organization changes in control and accounting systems. Davis et al (1986) examines if audit client with non-audit services result in knowledge spillover. They hypothesize a positive relation between audit and non-audit fees and used a sample size of 98 US clients. Their results suggest that non-audit firms providing audit and non-audit services is not a treat for their objectivity and like Simunic (1984) and Palmrose (1986), they found a positive association between audit and non-audit fees. Studies in other countries have also found a positive relationship between expenditures for audit and non-audit services. For example Barkess & Simnett (1994) investigate the relationship between audit and non-audit services in Australian setting. Their sample size existed of 500 publicly listed companies. Like the afore mentioned studies their study also indicated a positive relationship between expenditures for audit and non-audit service. Looking at studies conducted in UK, Ezzamel et al (1996; 2002) examines the relationship in a UK setting and they have showed the same results. Ezzamel et al. (1996) states that there is lack of information concerning non-audit fees that affect, directly as well indirectly, audit fees.

In contrast to the above studies, Abdel-khalik (1990) found no significant relationship between audit and non-audit services. He examined 84 US companies. Another contrary study is from O’Keefe et al (1994). His results did also not indicate significant relationship between audit and non-audit fees. Krishnan & Yu (2010) wielded a longer time period in the US and they found a strong and negative relationship between audit and non-audit fees. Table 1 summarizes some prior studies concerning audit and non-audit fees.

(9)

9 Table 1: Prior studies audit and non-audit fees.

Prior studies Sample Timeperiod

Relationship audit and non-audit fees Simunic (1984) 397 Public US companies December 1976-December

1977

+ Palmrose (1986) 298 US Public companies

October 1980-October 1981

+

Davis (1993) 98 US Companies +

Barkes & Simnet (1994)

500 US listed companies 1986- 1990 +

Ezzamel (1996) 361 UK listed companies 1 October 1992 - 30

September 1993

+

Ezzamel (2002) 193 UK listed companies 1994-1995 +

Firth (1997) 157 US listed companies 1991-1992 +

Abdel-khalik (1990) 84 US Companies 1987 No sig. relationship O'keefe (1994) 249 US companies 1989 No sig. relationship Krishnan & Yu (2010) 3110 US companies 2000-2006 -

Possible explanations for the positive association between audit and non-audit fees are identified by Solomon (1990). These explanations are based on the studies discussed above. The first explanation is that client who purchased both services could have been “problematic” firms that required non-audit services along with an extraordinary quantity of audit services (Solomon, 1990, p. 324). A second explanation is the changes in the client’s organization which may create need for additional audit effort. This can results in greater total audit fees. Another explanation is related to the economic characteristics. Solomon (1990) gives the example of audit services market is competitive but the market for audit services is a monopoly. In this case, “to obtain the services of the monopoly non-audit service supplier, the client may be willing to pay a higher non-audit fee than would be paid if only the audit services were purchased” (Solomon, 1990, p. 324). The last explanation refers to the internal dynamic of audit firms. Firm compensation schemes may create incentives for and audit partners may have to power the influence the otherwise arbitrary fee categorizations. In this case audit fees would appear to be greater, when in reality total fees paid by the client (i.e., for audit and non-audit service) could have fallen (Solomon, 1990, p. 325).

2.2. UK Setting

The United Kingdom (UK) consists of England, Wales, Scotland (who together make up Great Britain) and Northern Ireland. Companies listed on a stock exchange in the United Kingdom will be required to follow International Financial Reporting Standards in their consolidated financial statements starting in 2005 (www.iasplus.com). There are mandatory and advisory sources of generally accepted accounting principles (GAAP) in the UK. The mandatory sources are:

(10)

10  The Companies Act 2006

 Financial Reporting Standards (FRS) issued by the UK Accounting Standards Board (ASB)  Statements of Standard Accounting Practice (SSAPs) adopted by the UK Accounting

Standards Board

 UITF Abstracts issued by the Urgent Issues Task Force (UITF) of the UK Accounting Standards Board

 Listed companies must also comply with the reporting requirements set out in the Listing Rules of the London Stock Exchange, and AIM companies must comply with

The Accounting Standards Board (ASB), the Financial Reporting Review Panel (FRRP), and the Financial Reporting Council (FRC) together make up an independent private-sector organization whose purpose is to promote good financial reporting. The ASB and the FRRP are operational bodies.

During the financial crisis the audit sector got a lot of attention. Many companies that crashed at the begin of the financial crisis received an unqualified opinion from auditors. Most of the attention was focused on audit quality and confidence in accounting firm. In order to stimulate debates concerning changes in the audit practice the European Commission has released a document “Green Paper”. This paper discusses the role of the auditor, the independence issue and the audit market. A main issue is the auditor independence. The level and structure of the fees can affect the independence of the auditor. “Types of safeguards to be applied to mitigate or eliminate those threats include prohibitions, restrictions, other policies and procedures, and disclosure” (2006/43/EC). In the UK both audit and non-audit fees are disclosed in the company’s annual report. To avoid the independence threat, the EU requires a rotation of audit partners every seven years: ‘[…] the key audit

partner(s) responsible for carrying out a statutory audit rotate(s) from the audit engagement within a maximum period of seven years from the date of appointment and is/are allowed to participate in the audit of the audited entity again after a period of at least two years’ (European Parliament, 2006, p.

104). Audit-partner rotation is implemented in a number of countries, such as the United States, the United Kingdom, the Netherlands, Germany, Singapore, France, Spain, Japan, Taiwan and Australia (Chen et al., 2008; Chi et al., 2009). Regarding non-audit services, there are concerns about differences in the application of the EU 8th Company Law Directive by different European Union members. For example, in France it is forbidden to render non-audit services while it is allowed in UK (Green Paper, 2010/561/EU, p.12).

The discussed “Green paper” sets the basis for the comprehensive proposal. If the proposal will be approved, this will have radically changes for the audit market, especially for financial institutions and listed companies. Proposals that could have a major impact are according to KPMG (2012):

(11)

11  Mandatory audit firm rotation

 Mandatory tendering for appointing auditors  Restrictions on non-audit services

 Audit firm network meeting certain criteria would have to become audit-only within the EU  New requirements for audit committees and auditor reporting.

In the United Kingdom the Big Four companies present together the largest provider of audit services. The Big Four companies are PriceWaterhouseCoopers (PwC), KPMG, Deloitte and Ernst & Young.

Figure 2: FTSE 100: Market share of Big Four Auditors. Source: Telegraph, 2011

The Big Four firms audit 99% of companies listed on the London Stock Exchange with an average of auditors change every 48 years. The reason that the big four firms audit 99% of listed companies can be assigned to the difference between the Big Four and the mid-tier firms. According to audit committee chairs, finance directors and institutional investors there is a difference between the quality of staff, international coverage and reputation of mid-tier firms and Big Four firms (Oxera, 2006). The last decade there is a high degree of market concentration and the switching rates to other audit firms are low. Market concentration, increase of costs and changes in regulation may reasons that led to higher audit fees. According to the study of Oxera (2006) the larger firms are unlikely to employ a mid-tier firm, even if they were offered a substantial reduction in the audit fee (Oxera, 2006, p. 39). In 2010 the Big Four had a fee income of £7230 million (including audit and non-audit services) and a turnover of £7.16bn.

2.3. UK Financial crisis

This study examines whether economic conditions influences the relationship between audit and non-audit fees in the United Kingdom. Before the financial crisis occurred the economy was growing and investors and consumers had faith in the market. Extremely high fees are paid to auditors during the economic downturn (Sikka, 2009, p.868). After the financial crisis the confidence is affected and the

(12)

12 role of the auditor is questioned. It is critical to understand what kind of impact the financial crisis had brought to act more responsive in times of changes (good or bad). Taking a time period of good market conditions and bad market conditions an evaluation can be made of the relation and effect of audit and non-audit fees.

Figure 3: Economic Growth Europe. Source: Eurostat

Figure 3 highlights the economic growth of European countries. The economic growth of the UK was average compared to other countries but relatively stable. Before the financial crisis the inflation rate and the unemployment rate were relatively low. The credit crunch of August 2007 has led to the official recession in Britain in January 2009. The Office for National Statistics (ONS) announces that the gross domestic product (GDP) showed a fall of 1.5 per cent in the last three months of 2008 after a 0.6 drop in the previous quarter. This is the biggest quarter-on- quarter decline since 1980. This is the first recession since 1991 (Recession Report Britain) and has led to sharp falls industrial production.

Concerning the inflation rate, there is an increase since the financial crisis. In 2007 the inflation rate was 2.12 while a year after the inflation rate was 3.11%. The inflation has even arisen to 4.20% in 2011 (inflation rates, 2014). The August ONS figures show that the number of jobs fell by 271,000 in the three months to June 2009 (and by 573,000 over the year), the largest quarterly fall since modern records began in 1971.The employment rate for people of working age was 72.7 per cent for the three months to June 2009, down 0.9 from the previous quarter. Total employment now stands at just under 29 million people In countries like Britain and the United States, where labor market flexibility is relatively high, the fall in productivity growth is likely to be greater but the recovery swifter. This is because eventually the downturn leads the weakest firms to close(Recession Report Britain).

(13)

13

2.4. Determinants audit fees

2.4.1. Size of client

Prior studies showed that the size of client is one of the most important factors that affect audit fees (Hay et. al., 2006). Simunic (1980) provides an economic analysis of competition within the market for audit services. In order to do this, he developed a model which controls for factors affecting the audit fees. To measure the size of the client, total assets were used. The reason to use this as measure of size is: “the stock of assets seems more closely related to possible loss exposure than would an accounting flow measure, such as revenue, because defective financial statements which result in a lawsuit frequently involve some deficiency in asset valuation” (Simunic, 1980, p.172). The most studies that followed after Simunic (1980) also used total assets as a measure of size of client. Besides total assets, also revenues can be used (Hay, 2006, p.169).

The study of Simunic (1980) provided evidence that size is an important determinant of audit and that there is strong positive relationship between the size and the amount of audit fees. This is also confirmed by studies following the study of Simunic (Simunic, 1984; Francis & Stokes, 1986; Gerrald et.al. 1994; Al-Harshani, 2008). Simunic is one of the first studies that provided evidence concerning audit fees. Also he hypothesized that the relationship between size of client and audit fees is not linear. Not linear is meaning that the audit fee will increase in a lower rate as the client increases (Gerrald et.al., 1994, p.4). Also this is confirmed by other studies. Therefore, I will use the natural log of the client’s total assets in the models.

2.4.2. Complexity

Besides the size of client, it is proven that complexity is also an important determinant of audit fees. Chan et al. (1993) hypothesized that the increased effort where there is complexity will have a positive effect on the audit fees. The logical explanation for this is that more time and efforts are required from auditors to monitor organization’s activities (Simunic, 1980, p.172).According to the literature there is no direct measure of complexity, because complexity is a broad phenomenon that can be measured in a variety of ways. According to Mellet et al. (2007, p. 164) “proxies have to be used which broadly reflect the likely number of sub-systems in the accounting process, each of which have to investigated to form an audit opinion”. In prior studies focused on UK quoted companies different proxies were used. Ezzamel et al. (1996) examined the relationship between various categories of non-audit services and audit fees and used the number of subsidiaries, while Iyer and Iyer (1996) examined the effect of the Companies Act disclosure requirements on audit fee and non-audit fees and used the ratio of stock to total assets. The motivation for this is that it is hypothesized that stock is a relatively complex item and is susceptible to fraud (Firth, 1997). For this study I will use the number of subsidiaries and the number of foreign subsidiaries. Like the size of client, it is

(14)

14 hypothesized by Palmrose (1986, p.100) that there is no linear relationship between complexity and audit fees. She proved that the natural log of complexity is positively correlated with audit fee.

2.4.3. Risk

Audit risk is defined by Warren (1979, p. 66) as: The probability of issuing an inappropriate opinion on financial statements because material errors of irregularities, if they exist, will not be detected”. According to the literature there are three aspects of risk namely inherent risk, profitability and leverage.

The inherent risk is the probability that financial statements contain material misstatements without consideration of internal control (Eilifsen et al., 2010, p. 77). The study of Hay et al. (2006, p.170) proved that inherent risk is positively associated with audit fees. The motivation for this is that when an auditor suspect that the financial is misstated, more audit procedures are required. The most risky components of the balance are according to Simunic (1980) receivables and inventories. The inherent risk is usually measured, individually or in combination, by receivables and inventories divided by total assets. The study of Hay et al. (2006) indicates that the combination of receivables and inventories are used more often than individual.

The second risk that involves audit fees is profitability. Profitability reflects the extent to which the auditor may be exposed to loss in the event that a client is not financially viable (Simunic, 1980). If the performance of a company is not good, the audit takes more risk and demands higher audit fees. The study of Hay et al (2006, p. 170) states that “two variables are typically used to measure performance are a profitability ratio (usually net income divided by total assets) and a dummy variable for the existence of a loss”. Hay et al. (2006) expected that the relationship between audit fee and return on assets (ROA) will be negative. For the relationship between audit fees and loss they expected a positive sign. The results of their study have showed mixed results.

The third risk is related to leverage. Leverage measures the risk of a client failing, which potentially exposes the auditor to loss (Simunic, 1980). The most common proxies for leverage have been the ratio of debt to total assets and the quick ratio (Hay et al, 2006, p. 171. Prior studies showed a positive association between fees and leverage ratio and a negative association between fees and quick ratio (Gist, 1992; Fergusson et.al., 2003; Al-Harshani, 2008).

2.4.4. Size of auditor

Another determinant of audit fees is the size of auditor. According to Francis (1984, p.134), “the effect of audit firm size on audit prices is a complex function of competition in the market for audit

(15)

15 services, product differentiation, and scale economics to large firms”. Several studies investigated the effects of auditor size on audit fees.

The study of Palmrose (1986) tested if there is a systematic relation between audit firms sizes (both in terms of absolute size and relative shares) and audit fees. The results indicate a positive relation between audit fees and audit firm absolute size, treating the Big Eight as a homogeneous group of large firms. Palmrose (1986) did not found an association between audit fees and industry specialization. Another study that focused on the relationship between the auditor size and audit fees is Simunic (1980). In his study he divided the market into two segments: (i) with many smaller suppliers and (ii) Big Eight firm. The reason to divide the market is that the author expected to see price differences. He expected that the many smaller suppliers would me more competitive. However, the results were not significant. The author’s explanation for the finding was that “the Big Eight firms enjoy scale economies which are passed on as lower prices to auditee” (Simunic, 1980, p. 188). The study of Chan et al (1993, p. 781) argue that the product differentiation of large accounting firms is maintained by higher quality employees, who get higher salaries, which is reflected in audit fees. Firth (1985, p.28) suggested that the major accounting firms can spend more resources on training their employees. The size of auditor will be tested with Big Four/ Non-Big Four.

2.4.5. Type of industry

Simunic (1980, p.172) mentioned the industry of audited as a possible determinant of loss exposure. Pearson & Trompeter (1994, p.123) discussed a substantial difference between life and health insurance industries and the property and casualty insurance industries in relation to audit complexity. Stein et al. (1994) investigated the impact of industry differences in terms of number of hours of various type of professional labor required for audit service production. The results indicated a greater variation of audit efforts for financial institutions. Hay et al (2006, 9.175) argued that financial institutions and utilities have a different asset structure compared to other industries. In the sample size the financial institutions and utilities will be excluded.

2.4.6. Change of auditor

Change of auditor is another determinant that influences audit fees. Change of auditor can lead to possible fee reduction (low-balling). Low balling is defined by DeAngelo (1981, p.113) as: “the setting of the initial audit fee below the sum of audit start-up cost plus normal profits”. The study of Hay et al (2006, p. 176) provide two possible reasons for fee reduction. The first reason is that audit firms want to win new client. The other reason concerns the efficiency of service by audit firms. A new auditor can be more efficient. The study of Simon & Francis (1988) determine the presence and magnitude of audit fee price cutting. They used a sample size of 214 firms having changed auditors

(16)

16 over the period 1979-1984 and a control sample of 226 firms that did not change auditors over the same period. The results showed a significant fee reduction of about 24% during the year of initial engagement and 15% for each of the next two years. The study of Gregory and Collier (1996) examines the effects of auditor change on audit fees in Great-Britain. The results indicate that price cutting and subsequent price recovery also occur in Europe. Like the study of Simon and Francis (1988) the initial fee reduction is large and significant (around 22.4%). Also the result suggests that companies changing from non-Big Six to Big Six may experience an even greater reduction in fees. The most common proxy to measure the change of auditor in the literature is a dummy variable. Dummy variables indicates if auditor is new then there is probability of lower fee (Hay et al, 2006, p. 176)

2.4.7. Degree of competition

The last determinant discussed is the degree of competition. Several studies found evidence that the degree of competition is associated with audit fees. The results of most these studies indicate that audit fees decrease when the competition is high. Simunic (1980) argued that competition have an influence on audit fees, especially for larger firms. Maher et al. (1992) supports this result. The study of Maher et al. (1992) described the behavior of audit fees during a period of apparent increasing competition. The time period that was examined is from 1977 to 1981. They found evidence that audit fees decreased significantly during a period of increased competition in the audit market. In contrast, another study Iyer & Iyer (1996) examined the behavior of audit fees in the UK during a period of recent mergers among the Big 8 audit firms. The result of this study did not found evidence that despite increased concentration in the market for audit services that there was a significant increase in audit fees. Abdel-khalik (1990) examines the relationship between cost of auditor change and economic conditions in different regions in the US. He state that “an economic decline below expectations reduces demand for audit services and intensifies competition among audit firms. In contrast, regions expiring economic growth create more demand for the available services and provide opportunities for new entrants into the segment of audit market” (Abdel-khalik, 1990, p. 300).

2.5. Hypothesis development

This study investigates whether there is a positive relationship between audit and non-audit fees with the emphasis on economic conditions that may influence this relationship. Prior studies that investigate the association between audit and non-audit fees have shown mixed results. The studies of Simunic (1984), Palmrose (1986), Ezzamel et al. (1996) and Firth (2002) found evidence that there is a positive association between audit and non-audit fees. There are also studies that found no significant association (Abdel-Khalik 1990; O’Keefe et al. 1994). Contrary to most of the studies, the

(17)

17 study of Krishnan & Yu (2010) found a strong and negative association between audit and non-audit fees.

Most of the discussed studies are conducted in the US (Simunic, 1984; Palmrose, 1668a; Abdel-khalik, 1990). Relatively little research has focused on the UK market. The US and UK market for audit and non-audit fees differ in the accounting and audit environment (Gwilliam, 1987). Therefore the general assumption that US research is automatically transferable to a UK setting is not correct. One difference between the US and UK market is that provision of non-audit services in US is significantly less extensive than in the UK (Gwilliam, 1987). Another difference suggested by Ezzamel et al. (1996) is the differences in the nature of the non-audit services market. UK companies are more successful in dominating a position in taxation advice.

The studies of Ezzamel et al. (1996, 2002) are interesting for this study. Their study is also conducted in the UK, but their setting was different. Another interesting study is the study of Krishnan & Yu (2010) who showed, contrary to most other studies, a negative relationship between audit and non-audit fees. The first study of Ezzamel et al. (1996) used a sample of 314 quoted companies (excluding financial sector companies) and examined also the relationship between audit and non-audit fees. The time period covered one year. The results were consistent with US studies. Also the study revealed that non-audit service provided a main source of income for audit firms, calculated to be about 87% of total audit fees (Ezzamel et.al, 1996, p.13). However, covering one time period may influence the results. The present study investigates different economic conditions. Focusing on one year, where the state of economy is not covered in the study, we cannot draw any conclusion. The second study of Ezzamel et al. (2002) used a sample of 193 companies and used a classification of non-audit fees. However, there is no universal classification of types of non-audit services to adopt. This makes the results less reliable for other studies. This study also covers one year. Krishnan & Yu (2010) is distinctive from other studies by using a larger sample size. The smaller samples used by other studies may be biased. The larger sample size explains the divergent result of a negative relationship between audit and non-audit fees.

Possible explanation for a positive association is provided by Solomon (2010). The most of the studies that are discussed above had a time period of one year and did not have a large sample size. Firth (1997) suggested that the small sample size and short time period may influences the results. This seems plausible looking at the study of Krishnan & Yu (2010). Their sample size is way bigger and they cover a time period of six years. The results of their study showed a negative correlation between audit and non-audit services. This results leads to the suspicion that results examining a short period are less reliable than longer time periods. Therefore, more research based on longer time period is needed to explain the association between audit and non-audit fees. This study will distinguish her

(18)

18 and will have a contribution to the exiting literature due to the larger sample size and time period that will be used. Based on prior literature, I expect a positive association between audit and non-audit fees in the UK.

H1: The relationship between audit and non-audit fees is positive in the United Kingdom

The focus on this paper will be on macroeconomic conditions. Some papers provide indications that economic downturn may affect the relationship between audit and non-audit fees (Abdel-khalik 1990; Palmrose 1986; Firth 2002). Looking at the demand for products, we can say something about the state of the economy. When the economy is doing well, there is an increase in the demands for products. The financial crisis is an example of economic downturn. There was a sharp decline in demands for product, which lead to a sharp decline in the liquidity of companies. Companies did not have another choice than to restructure their operation. By lack of experience for restructuring the business, we can assume that the demand for non-audit fees increased. Therefore I expect that in times of crisis, there will be in increase in the demand for non-audit fees.

The study of Firth (2002) focused on specific events that affect non-audit services. In times of specific events, like mergers, acquisitions and business restructuring, increase, the demand for non-audit services increased. Due to changing accounting and information systems, more time and effort is required by audit firms. This is also confirmed by Palmrose (1986). She stated that purchasing more non-audit fees because of changes in the organization will also require more effort from the auditor to conduct the audit.

Studies suggest that competition is the critical factor that influencing the price of audit and non-audit services firms (e.g. Simunic, 1984; Maher et.al. 1992; Pearson & Trompeter, 1994). Maher et al (1992) examined audit fees in a period of apparent increasing competition in the market. The study has found a significant decrease in audit fees. Abdel-khalik (1990) investigates the relationship between completion on audit market and economic conditions. He states that during an economic downturn the clients are likely to be able to negotiate for a lower price for service.

Assuming that in times of mergers, acquisitions and business restructuring the demand for non-audit services increased I expect that also in economic downturn (also involve of business restructuring) there will be an increase in the demand for non-audit services. Although this is likely the case, we have to keep in mind that reduction in solvency of companies may prevent companies from buying non-audit services. Besides the many business restructuring during the economic downturn, we can also assume based on the study of Maher (1992), that the degree of competition is higher during the crisis. I hypothesize that there is a positive relationship between audit and non-audit fees and that this

(19)

19 relationship will be stronger in times of economic downturn. Therefore I expect that the regression for the three different time periods will have different results. Specifically, I expect that the time period 2008-2009 will show different results and that the ratio of non-audit fees increase and the ratio of audit fee will decrease the most. The above discussion leads to hypothesis 2, 3 and 4.

H2: The changed economic conditions affect the relationship between audit and non-audit fee H3: The financial crisis affects the ratio of audit fees.

H4: The financial crisis affects the ratio of non-audit fees

As described above the economic turndown may affect the relationship between audit and non-audit services. However, it is difficult to assign possible changes to only economic conditions, while there is variety of factors which can influences audit fees. This study investigates whether the relationship between fees and non-audit fees is positive. Second, I will investigate whether the relationship between fees and fee ratios changes and if economic conditions has an influence on this relationship. By examining fees during a longer period, the results of this study will contribute to prior literature.

(20)

20

3. Methodology and sample selection

In this section the sample selection and the dependent and explanatory variables are discussed. The following paragraphs also specify the control variables and conclude with estimations for these variables. The present study will be based on a quantitative approach. Hereby I will examine the relationship between audit and non-audit-fees. The analyses are based on data of publicly traded firms operating in the UK. The firms are selected from the FTSE All Share Index(LFTALLSH). The reason to choose UK is because the mandatory disclosure of both audits as well non-audit fees in the annual report. I obtained the data necessary for running the regression from DataStream. In DataStream data about the number of UK subsidiaries and foreign subsidiaries is not available. The database Amadeus is used to get information about the subsidiaries.

In order to the test the relationship between audit and non-audit fees in UK (H1) and test whether economic conditions have an impact (H2) on audit and non-audit fees, four samples are used. I have three samples of two years and the fourth sample is the total of the three samples together. For all three samples, the same companies are used. Thus, in one subsample the same company occurs twice and six times in total. Companies that not occur each year are deleted from the sample to provide a better comparison between the time periods. The three time periods are:

 Favorable period: 2006-2007  Economic downturn: 2008-2009  Post-crisis period: 2010-2011

The total number of identified companies over the 6 year is 3.768. Each year there are 628 firm observations. In accordance with previous studies, financial institutions and utilities are excluded because of their different assets structure. The remaining industries are: basic materials, consumer goods, consumer services, health care, industrials, oil & gas, technology and telecommunications. After deleting the financial institutions and utilities, also companies with a fiscal year that did not corresponds to calendar year are deleted from the sample. Thus, only companies which fiscal year is between January and December are included. This is done to provide more consistent results. As last, companies with missing variables and companies that did not occur every year are removed from the sample. In the end, all three samples exist out of 170 observations (85 per year) and the total number of observations over the six years is 5102. Table 2 summarizes the realization of the final sample.

2_{The company BRITISH AMERICAN TOBACCO P.L.C. showed extremely high value in audit fees in the}

year 2007. This company is considered as an outlier and is deleted in all six years to provide more consistent results.

(21)

21 Table 2: Sample Size

Favorable Economic downturn Post-crisis

2006 2007 2008 2009 2010 2011 Total

Initial sample 628 628 628 628 628 628 3768

Exclude: financial institutions, utilities and odd fiscal year -465 -460 -455 -457 -454 -449 -2740 Sample after exclusion 163 168 173 171 174 179 1028 Exclude: Missing variables and companies that do not occur every year -77 -82 -87 -85 -88 -93 -512 Sample 86 86 86 86 86 86 516 Sample periods 172 172 172 516 Exclude: outliers -2 -2 -2 -6 Total Sample 170 170 170 510

I collected data on audit fees, non-audit fees, total assets, receivables, inventory, parent auditor, total debt, industry group, UK subsidiaries and foreign subsidiaries. Based on this data, other variables are composed. I will explain these variables together with the models.

As before mentioned the financial institutions and utilities are excluded because they have a different asset structure. The different asset structure may affect the comparisons. In the next figure the sample distribution by industry is provided. Industrials are the most present in one subsample (72). The smallest industries are the telecommunications and technology (4, 6 respectively). In figure 4 the sample distribution by industry per period is provided.

Figure 4: Sample distribution by industry per period

0 10 20 30 40 50 60 70 80

(22)

22 Figure 5 highlights the audit fees and non-audit fees that are paid to auditor firms by companies. As illustrated there is an increase in both audit as well non audit fees for the period 2008-2009. This is in line with prior literature which stated that during the economic downturn large amounts of audit and non-audit fees are paid to audit firms. This overview also confirms the hypothesis that the crisis influences the relation of audit fees and non-audit fees.

Figure 5: Overview averages audit and non-audit fees

The dependent variable in this study is the natural log of audit fees. The explanatory variables are client size, complexity, risk and auditor size. With the collected data I can add these variables in the regression.

Size of client will be measured with the natural log of total assets. Prior studies found evidence that

there is strong positive relationship between the size and the amount of audit fees. However, it is proven that the relationship between the client size and audit fees is not linear. Therefore I will use the natural log of the client’s total assets.

For the complexity I collected data of the number of UK subsidiaries and the number of foreign subsidiaries. Palmrose (1986) found evidence that there is no linear relationship between complexity and audit fees. I will use the natural log of the UK and foreign subsidiaries. The study of Palmrose (1986) found that complexity is positively associated with audit fees.

To measure risk I will use four proxies. According to the literature the most risky components of the balance sheet are receivables and inventories. I will use these components and divide them by total assets. Second, I will look at the return on assets (ROA). The ROA is the return on assets calculated as the net income deflated by total assets. Poor performance of a company implies increased risk for auditors, which implies higher audit fees. Third, a dummy variable is used for the existence of loss. Loss and audit fees are positively correlated. Thus, more loss means higher audit fees. At least, I look at the leverage. Clients with a higher leverage ratio are likely to be more risky and will have higher cost for audit.

0 1.000.000 2.000.000 3.000.000 4.000.000 5.000.000 2006-2007 2008-2009 2010-2011 Audit Fees Non-audit Fees

(23)

23 The size of auditor will be measured with the dummy variable Big 4/ Not Big 4. Figure 6 provides an overview of the explanatory variables.

Figure 6: Key determinants audit fees

The initial model to test the impact of different factors on audit fees is provided by Simunic (1980).  AUDIT = βo + β1ClientSize+ β2 Complexity+ β3 Risk+ β4 AuditorSize + β5NonaudFee+ e In this study the models looks slightly different. I will use the basic model to test H1 and H2. The following model includes all years namely from 2006 to 2011.

 LNAuditFees2006-2011= βo+ β1LNNonAuditFees2006-2011+ β2LNTotAssets2006-2011+

β3LNUKSUBS2006-2011+ β4LNFSUBS2006-2011+ β5INVREC2006-2011 + β6ROA2006-2011 +

β7LEV2006-2011+ β8LOSS2006-2011 + β9AuditFirmSize2006-2011 + β10ROT2006-2011 + e

Because I want to test several economic conditions this model will be applied to three different time settings. The following models will be used to test all hypotheses.

The model for time period 2006-2007:

 LNAuditFees2006-2007= βo+ β1LNNonAuditFees2006-2007+ β2LNTotAssets2006-2007+

β3LNUKSUBS2006-2007+ β4LNFSUBS2006-2007+ β5INVREC2006-2007 + β6ROA2006-2007 +

β7LEV2006-2007+ β8LOSS2006-2007 + β9 AuditFirmSize2006-2007 + β10ROT2006-2007+ e Size of client total assets Complexity number of UK subsidiaries number of foreign subsidiaries Risk (receivables+i nventories)/to tal assets ROA Existenceof losses Leverage Size of auditor Big 4/ Not Big 4

(24)

24 The model for time period 2008-2009:

 LNAuditFees2008-2009= βo+ β1LNNonAuditFees + β2LNTotAssets2008-2009+ β3LNUKSUBS 2008-2009+ β4LNFSUBS2008-2009+ β5INVREC2008-2009 + β6ROA2008-2009 + β7LEV2008-2009+

β8LOSS2008-2009 + β9 AuditfirmsSize2008-2009 + β10ROT2008-2009 + e

The model for time period 2010-2011:

 LNAuditFees2010-2011= βo+β1LNNonAuditFees2010-2011+β2LNTotAssets

2010-2011+β3LNUKSUBS2010-2011+β4LNFSUBS2010-2011+β5INVREC2010-2011+β6ROA2010-2011 +

β7LEV2010-2011+β8LOSS2010-2011+β9AuditFirmSize2010-2011+β10ROT2010-2011 + e

Table 3: Variables used in this study

Variable Explanation Type Expected effect

Audit Audit fees (GBP)

Non-Audit Non-audit fees (GBP)

LNAuditFees Natural log of company's audit fees Dependent LNNonAuditFees Natural log of company’s non-audit fees Explanatory + LNTotAssets Natural log of total assets Explanatory + LNUKSUBS Natural log of the number UK subsidiaries Explanatory + LNFSUBS Natural log of the number of foreign subsidiaries Explanatory + INVREC Inventory plus accounts receivable, divided by total assets Explanatory + ROA _{Return on assets, calculated as operating income divided by}

total assets Explanatory -

LEV _{Total debt divided by total assets} _Explanatory _- LOSS _{A dummy variable is coded 1 if the net profit is negative in}

previous year,

Dummy -

_{otherwise 0}

AuditFirmSize _{A dummy variable is coded 1 if auditor is a Big 4 firm,} otherwise 0

Dummy +

ROT(rotation) _{A dummy variable is coded 1 if the auditor’s engagement is}

the initial two years, otherwise 0

Dummy +

(25)

25

4. Empirical results

4.1. Descriptive statistics

Table 4 summarizes the statistics for the selected variables for the different time periods. The most important variables are the natural log of audit and non-audit fees. The highest value of the dependent variable audit fee logged in the crisis period (2007-2008). The mean and median values are 14.208 and 14.105 (see table 4). The lowest audit fee value (logged) is in the favorable period (2006-2007). The mean and median are 13.944 and 13.750 respectively (see table 4).

The independent variable non-audit fee (logged) shows the same pattern for audit fees. The mean and median during the crisis are 13.285 and 13.585. However, the lowest value of non-audit fees is in the post-crisis period with a mean of 13.093 and median 13.340. This results indicate that, like predicted, the audit and non-audit fees increased during the economic downturn and starts to decrease after the worse part of the crisis. After the financial crisis, both audit fees as well non audit fees decreases. However, non-audit fees decrease more rapid. An explanation for this is that during the financial crisis, there was a higher demand for non-audit fees because of the higher competition. Also like stated in the literature review, in times of economic downturn, companies have to reorganize and have a higher need for non-audit serves due their own lack of experience. Thus, after the reorganizations the demand for non-audit services declined.

The mean and median of logged total assets show a steadily growing during the tested periods. The highest value is in the post-crisis period with a mean of 14.772 and median of 14.604 and the lowest value in the favorable period with a mean of 14.343 and median 14.170. The mean (logged) of UK subsidiaries is the same before and during the economic downturn and is increasing in the post-crisis period. The mean (logged) of foreign subsidiaries is the same in the favorable and crisis period and is decreasing in the post-crisis period.

The proxy to measure risk INVREC was the highest in the favorable period with a mean of 0.292 and median of 0.182 and the lowest during the economic downturn 0.267 and 0.276 respectively. The mean of the return of assets (ROA) was highest in the favorable period (0.09) and then declined to 0.05 during the economic downturn and then increased again to 0.07 in the post-crisis period. The mean leverage was highest (0.22) during the economic downturn but does not differ much from the other periods (0.21 and 0.18).

The number of companies which reported negative profit in previous year was highest (0.09) after the financial crisis and the lowest in the favorable period. The most companies registered on the FTSE All Share Index have a Big Four as auditor. Before the financial crisis, 98% was audited by a Big Four

(26)

26 Companies. The economic downturn did not have an influence and this percentage remained the same. Also auditor rotation did not change during the six years. The variable ROT is constant and does not have an influence on the mode. Therefore the variables ROT will be excluded from the correlation analysis and the regressions.

In overall, the most important variable of interest LNNonAuditFees shows the pattern like expected. The audit fees and non-audit fees are high in the favorable periods, drops during the economic crisis. Non-audit fees declines less than audit fees. Also the variable LNTotAssets shows the same pattern. The control variables show mixed results. To draw any conclusion, the correlations and regression will be analyzed first. Table 4 provides the descriptive statistics for all periods.

(27)

27 Table 4: Descriptive statistics

Time

period LNAuditFees LNNonAuditFees LNTotAssets LNUKSUBS LNFSUBS INVREC ROA LEV LOSS AuditFirmSize Rot

2006-2011 Mean 14.116 13.200 14.594 2.053 3.544 0.280 0.07 0.200 0.078 0.976 1 Median 14.000 13.575 14.415 1.370 3.600 0 0 0 0 1 1 SD 13.646 2.433 1.466 2.212 1.582 0.182 0.089 0.155 0.269 0.152 0 Min 10.340 0 11.720 -0.370 0 0 0 0 0 0 1 Max 17.860 18.490 19.390 14.800 7.830 1 1 1 1 1 1 2006-2007 Mean 13.944 13.221 14.343 1.522 3.560 0.292 0.092 0.207 0.059 0.976 1 Median 13.750 13.590 14.170 1.140 3.600 0 0 0 0 1 1 SD 1.426 2.575 1.462 2.272 1.570 0.182 0.085 0.169 0.236 0.152 0 Min 10.370 1 11.720 -0.370 0.690 0 0 0 0 0 1 Max 17.790 17.860 19.020 14.190 7.830 1 1 1 1 1 1 2008-2009 Mean 14.208 13.285 14.667 1.581 3.560 0.267 0.050 0.221 0.000 0.976 1 Median 14.105 13.585 14.525 1.160 3.600 0 0 0 0 1 1 SD 1.337 2.288 1.452 2.438 1.570 0.176 0.099 0.157 0.276 0.152 0 Min 10.340 0.330 12.370 -0.370 0.690 0 1 0 0 0 1 Max 17.860 18.490 19.120 14.800 7.830 1 0 1 1 1 1 2010-2011 Mean 14.197 13.093 14.772 3.055 3.510 0.276 0.007 0.183 0.094 0.976 1 Median 14.050 13.340 14.604 3.040 3.597 0 0 0 0 1 1 SD 1.320 2.439 1.459 1.452 1.613 0.175 0.079 0.135 0.293 0.152 0 Min 10.370 0 12.440 0.000 0 0 0 0 0 0 1 Max 17.800 17.470 19.390 6.240 7.830 1 1 1 1 1 1

(28)

28 In order to test the fee ratios during the different time periods, I also used the same three samples. Descriptive statistics of the ratios are illustrated in table 5. The audit fee ratio is steadily growing during the different economic conditions, and non-audit fees are steadily increasing in the same proportion. In the next chapter, I will test whether the difference are significant different. Table 5: Audit and non-audit fee during different time periods

Period Audit fee ratio Non-audit fee ratio

Mean Std. Dev. Mean Std. Dev

Favorable 0,58 0.239 0,42 0.239

Economic Downturn 0,64 0.216 0,36 0.216

Post-crisis 0,66 0.207 0,34 0.207

4.2. Additional descriptive statistics industries

Like mentioned in the previous section, financial institutions and utilities are deleted. The remaining industries are basic materials, consumer goods, consumer services, health care, industrials, oil & gas, technology and telecommunications. Industrials are the most present (72) and telecommunications (4) and technology (4) is the least present in the sample size. I have compared the industries per time period to see how much they pay more or less than the previous period.

Descriptive statistics of audit fees per industry (table 6) shows that in het favorable period health care and oil & gas spent the most on audit fees. The industry technology and telecommunications spent the least on audit fees. In the crisis period no longer health care and oil & gas represents the biggest group, but basic materials. They pay 65% more on audit fees than in the favorable period. Health care shows a major decline (-30%) between the two periods. Telecommunications represent still the smaller group, but are relatively stable (-1%). Looking at the change between the crisis and the post-crisis, basic materials fluctuates again the most (-49%). Also technology shows a major decline, representing the smallest group. Descriptive statistics of non-audit fees per industry (table 7) shows that basic materials are the biggest group in the favorable period. Industrials represent the smallest group. In the crisis period basic materials is again the biggest group. Also this group is fluctuating the most. There is an increase of 127 percent. The fluctuation for the other industry is also big (technology -39%, telecommunications -37%). Between the period crisis and post-crisis basic materials shows again a major decline, together with oil & gas and technology.

Overall, audit fees are increasing with 9% from the favorable period to the crisis, and are then decreasing with -14% in the post-crisis period. Non-audit fees show first an increase of 28% and then a decrease of -37%. Basic materials fluctuate the most between the periods. Because the small sample, the fluctuations of the industries cannot be statistically tested. From an economy point of view, table 6 and 7 provided an overview of the sensitivity of industries.

(29)

29 Table 6: Change in percentage audit fees per industry

Audit fees 2006-2007 2008-2009 2010-2011 Basic materials 6,687,750 65% 11,036,417 -49% 5,623,720 Consumer goods 2,753,278 26% 3,480,486 -8% 3,192,111 Consumer services 2,859,004 0% 2,863,179 4% 2,965,417 Health care 8,144,684 -30% 5,691,200 7% 6,077,200 Industrials 1,696,220 23% 2,084,332 5% 2,195,911 Oil & gas 8,391,519 1% 8,484,840 6% 8,973,373

Technology 1,162,806 -16% 977,779 -26% 724,333

Telecommunications 1,370,293 -1% 1,350,000 0% 1,350,000

Total 33,065,553 9% 35,968,232 -14% 31,102,065

Table 7: Change in percentage non-audit fees per industry Non-audit fees 2006-2007 2008-2009 2010-2011 Basic materials 6,330,000 127% 14,352,583 -50% 7,218,333 Consumer goods 1,248,889 8% 1,345,389 -16% 1,123,667 Consumer services 2,591,671 -6% 2,436,779 -5% 2,307,975 Health care 3,093,915 12% 3,472,700 -26% 2,582,600 Industrials 914,463 -1% 905,235 -9% 823,350

Oil & gas 4,969,892 -16% 4,182,549 -66% 1,436,228

Technology 1,502,529 -39% 923,239 -12% 808,000

Telecommunications 1,938,529 -37% 1,225,000 63% 2,000,000

Total 22,589,888 28% 28,843,473 -37% 18,300,153

4.3. Correlation analysis

The correlation analysis used to measure the strength of the relationship between distributed variables. I will use the Pearson Correlation Coefficient. The coefficient can take a value between -1 and 1. A value near to 1 is a sign of strong positive association. A value near to -1 reveals a strong negative linear correlation. By using the Pearson Correlation analysis, I can examine the extent of multicollinearity between independent variables. This is important because the tested independence variables may cause bias in the interpretation of the regression. According to Judge et al. (1988, p. 868), correlation below 0.8 should not be too biased to multicollinearity.

(30)

30 Table 8: Correlation matrix total period

Table 9: Correlation matrix favorable period

Table 10: Correlation matrix economic downturn period

LNAuditFees

LNNonAudit

Fees LNTotAssets LNUKSUBS LNFSUBS INVREC ROA LEV LOSS

AuditFirm Size LNAuditFees 1 LNNonAuditFees ,499** 1 LNTotAssets ,796** ,510** 1 LNUKSUB ,241** ,085 ,291** 1 LNFSUB ,624** ,287** ,490** ,276** 1 INVREC -,238** -,199** -,173** -,047 -,235** 1 ROA ,011 -,044 -,102* ,039 ,080 -,061 1 LEV ,132** ,140** ,169** ,000 ,201** -,235** -,151** 1 LOSS -,051 ,023 -,036 -,003 -,137** ,081 -,177** -,013 1 AuditFirmSize ,099* ,215** ,155** ,047 ,021 ,142** ,015 ,172** -,003 1 **. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).

Correlations 2006-2011

LNAuditFees

LNNonAudit

Fees LNTotAssets LNUKSUB LNFSUB INVREC ROA LEV LOSS

Auditfirm Size LNAuditFees 1 LNNonAuditFees ,339** ₁ LNTotAssets ,748** _,401** ₁ LNUKSUB ,395** _,027 _,405** ₁ LNFSUB ,578** ,201* ,481** ,467** 1 INVREC -,175* -,101 -,117 -,052 -,243** 1 ROA -,034 ,007 -,107 -,097 ,036 -,043 1 LEV ,056 ,066 ,107 ,079 ,147 -,235** -,077 1 LOSS ,003 ,078 -,111 -,068 -,198* -,062 -,044 -,162* 1 AuditFirmSize ,082 ,361** ,205** -,001 ,022 ,141 -,133 ,151* ,039 1 **. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed).

LNAuditFees

LNNonAudit

Fees LNTotAssets LNUKSUBS LNFSUBS INVREC ROA LEV LOSS

AuditFirm Size LNAuditFees 1 LNNonAuditFees ,688** 1 ,817** ,662** 1 LNUKSUBS ,372** ,259** ,383** 1 LNFSUBS ,666** ,410** ,518** ,467** 1 -,255** -,309** -,161* -,019 -,243** 1 ROA ,067 -,102 -,099 -,048 ,095 -,223** 1 LEV ,147 ,230** ,224** ,105 ,183* -,188* -,233** 1 LOSS -,190* -,072 -,023 -,148 -,256** ,295** -,204** ,025 1 AuditFirmSize ,111 ,104 ,160* -,001 ,022 ,144 ,019 ,195* ,047 1 **. Correlation is significant at the 0.01 level (2-tailed).

*. Correlation is significant at the 0.05 level (2-tailed). INVREC

Effect of different economic conditions on audit and non-audit fees : evidence from the United Kingdom

Amsterdam Business School