The effect of industry specialization & audit tenure on audit quality

(1)

Amsterdam Business School

The effect of industry specialization & audit tenure on audit quality.

Student:

Name: Mattijs Mans Student number: 6129986 Study:

MSc Accountancy & Control, specialization Accountancy 2013/2014 Faculty of Economics and Business, University of Amsterdam Date: 23-6-2014

Supervisor: Dhr dr. Jeroen van Raak

(2)

2

Acknowledgements

The goal of writing this thesis is to obtain a Master of Science degree in Accounting & Control at the Amsterdam Business School. First of all I would like to thank my supervisor, Jeroen van Raak, who gave my guidance and support during the process. Furthermore I would to thank my employer KPMG, which gave me the opportunity, resource and time in this busy period to complete the master thesis. Beside that I would like to thank my family, friends and colleagues for support and understanding of missing another evening at the bar. A special thanks to Robert van Huizen for helping me out with all sort of question during the process.

(3)

3

Abstract

The effect of industry specialization, audit tenure on audit quality

There is an ongoing debate about the role of the auditor recently extended by the discussing of mandatory audit firm rotation due to accounting scandals during the beginning of the 21st century. Prior research suggest that industry specialization is associated with higher audit quality and audit quality is likely to increase with audit firm tenure due to a dominant learning effect.

I am combining those findings and examine whether industry specialists have the same learning effect over the years. In this master thesis I used a large sample of U.S. Firms to determine if industry specialization, measured at 5 different ways, and audit tenure effects the audit quality significantly. I examine the effects on audit quality using the amount of

discretionary accruals as a proxy for audit quality. A high amount of discretionary accruals is used as an indicator of low audit quality.

I found that auditor industry specialist has positive effects on audit quality when auditing a firm for more than 7 years. Therefore it could be stated that auditor industry specialist has a less dominant bonding effect for instance due to the greater reputation loss of the industry specialist.

The findings have to be interpreted carefully because of the limitations in this study. It could be possible that other ways to measure audit quality would have led to other results. Next to this, the United States results cannot be generalized because of notable differences in the institutional backgrounds and the power of enforcement in other countries.

(4)

4

1 Introduction

The role of auditing in ensuring the quality of corporate earnings have been under considerable scrutiny due to major accounting scandals and the collapse of Enron in the US (Browning and Weil, 2002) and in the Netherlands by incidents with Vestia, SNS Reaal and Ballast Nedam. Every time the question is addressed: where were the auditors? Therefore, there is an ongoing debate about the role of the auditor. The discussion focuses mainly on the independence of the auditor. For instance, the discussion about mandatory audit firm rotation by the European commission (EU Green Paper 2010). Beside the EU the Public Company Accounting Oversight Board (PCAOB) is reviewing the idea of mandatory audit firm rotation. DeAngelo (1981a) describes two main variables to determine audit quality: Auditor independence and auditor competence. The main debate is about auditor independence but competence is probably at least as important since an auditor will never report a breach in the financial reporting if he does not find the breach. Industry specialization could improve the competence of the auditor due to the fact that having more clients in a specific industry could lead to an accelerated learning curve which could have a positive effect on audit quality.

Audit quality differences result in variation in credibility offered by auditors, and in the earnings quality of their clients. Since auditor quality is indefinitely and inherently unobservable, there is no single auditor characteristic that can be used to proxy it (Balsam et al. 2003). For a couple of years researchers have used the auditor brand name to proxy for audit quality and examined the association between brand name and earnings quality (Becket et al. 1998; Reynolds and Francis 2000). Other researchers (Craswell et al 1995; Beasley and Petroni 2001) have hypothesized that, in addition to brand name, an auditor’s industry specialization contributes positively to the credibility offered by the auditor.

Balsam et al. (2003) extend this literature by comparing the earnings quality of clients of industry specialist and non specialist auditors. Reichelt and Wang (2009) found that auditors who are both national and city-specific industry specialists have clients with the lowest abnormal accruals, suggesting that joint national and city-specific industry specialists have the highest audit quality. They used cross-sectional measures, these measures indicate an industry specialist as an auditor that has a lot of clients in an industry during a single year. The studies of Balsam et al (2003) and Reichelt and Wang (2009) argue that having a lot of clients in an industry in a specific year increases expertise and they find indeed a better audit quality for clients in this industry during that year. These studies deny the theory of learning (longitudinal effect). Several studies

(7)

7

suggest that auditors became more experienced over time due to a learning effect (See Simunic 1984, Morgan and Stocken 1998; King and Schwartz 1999; Solomon, Shields, and Whittington 1991; Low 2004). Most of these studies use the years of client experience to proxy the learning effect. The goal of this study is to investigate whether the experience an auditor gained in a specific year will affect the quality of the audit services in that industry even stronger in the following years. Therefore the research question is:

Does the industry specialization affects the audit quality?

The contribution of this paper is that by combining the theory of learning and industry specialization will lead to a better evaluation of the choice of an auditor and the credibility of his audit opinion.

This paper is composed as follows. In Section 2, I discuss the prior literature and definitions used for this study which lead to the development of the hypotheses. The sample selection and research design is discussed in Section 3. In Section 4 the descriptive statistics and results are analyzed. In section 5 I come to my conclusion.

(8)

8

2 Literature review

This chapter provides a review of existing literature to frame the research study. I will first explain what audit quality means and how it can be measured, secondly I will describe audit tenure and learning theories. In section 3 I will focus on the concept of industry specialist. Finally the hypotheses are developed and discussed.

2.1 Audit quality

DeAngelo (1981a) defines audit quality as the joint probability for an auditor to discover a breach (competence) and report the discovered breach (independence).

Figure 1 presents a framework for placing the current study within the existing research. This line of research examines whether audit quality is positively associated with financial reporting quality. Audit quality is proxied by quality of financial reporting which could be measured by using various variables (see figure 1). Auditor industry specialization is positively associated with quality of financial reporting.

source: Balsam(2003)

The initial studies in auditing focus on the association between auditor brand name and financial reporting quality measured by one of the items listed right in figure 1.

More recent literature argues that, in addition to brand name, an industry specialist in comparison with a non specialist offers a higher level of audit quality. (e.g., Craswell et al. 1995; Beasley and Petroni 2001). Recent literature links industry specialization with financial reporting

(9)

9

quality. For example Carcello and Nagy (2002) found evidence that clients of industry specialists have a smaller chance to be involved in SEC enforcement actions. Other research suggest that earnings of clients of industry specialists have a more accurate predictability of future cash flows than those of non specialists (Gramling et al., 2001)

In terms of Figure 1, I focus (similar to Balsam et al, 2003) on whether audit quality, as measured by industry specialization is associated with earnings quality. As described in many papers, management has an incentive to manage earnings (e.g. Jensen and Meckling, 1976). These incentives arise out of explicit and implicit contracts that link outcomes of interest to management (e.g, managerial compensation to reported earnings). The quality of the auditor hired by the firm is one factor that restricts the extent of earnings management. Defond et al. (2005) suggest that earnings management not detected by the auditor could be an indicator of low audit quality. Auditors have more control over discretionary accruals. That is, because of the discretionary accruals involve greater subjectivity (Mascarenhas et al., 2010).

To measure the amount of discretionary accruals the modified Jones model (1991) is commonly used. When using the Modified Jones model, total accruals can be estimated as follows: ‘Total accruals = net income – operational cash flow’. The total accruals can be divided into discretionary accruals and non-discretionary accruals. The proportion of total accruals unexplained by normal operating activities is discretionary accruals (Jackson et al., 2008). The normal accruals can be estimated as a function of the change in revenue and the level of propert, plant and equipment by using the modified Jones model 1991 (Becker et al., 1998). The amount of discretionary accruals indicates if the financial statement reflects the company’s ‘true’ operatin results and financial condition or not. A high amount of discretionary accruals is used as an indicator for lower earnings and audit quality, since it indicates that management is able to adjust earnings numbers in their best interest (Chen et al., 2008) .

Dechow et al. (1995) modified the original Jones model (1991) to detect earnings

management. ‘The modification is designed to eliminate the conjectured tendency of the Jones model to measure discretionary accruals with error when discretion is exercised over revenues. [….] The only adjustment to the original Jones model is that the change in revenues is adjusted for the change in receivables in the event period’ (Dechow et al., 1995, p.199) During this study I will also use the modified Jones model since Dechow et al. (1995) have tested several models for detecting earnings management and found that this modified Jones model had the most power to detect earnings management.

(10)

10

2.2 Auditor tenure and learning theory

Auditors accumulate client-specific knowledge as they perform audit engagements; their belief about clients are updated each period and become more precise over time. This can be named as the learning effect (Beck and Wu 2006). The learning effect has a favorable impact on audit quality (See Simunic 1984, Morgan and Stocken 1998; King and Schwartz 1999; Solomon, Shields, and Whittington 1991; Low 2004). Auditor tenure is defined as the number of years an auditor is retained by the firm (Myers et al., 2003). During the first years of the auditor’s tenure the learning effect occurs, where audit quality increases as the auditor obtains more

understanding of the client’s financial reporting system, business, industry and internal controls (Brooks et al., 2011). This effect is consistent with the ‘learning curve’ effect (Yelle 1979; Chen and Manes 1985) where new information is acquired after each successive audit engagement at a diminishing rate, until a maximum amount of information is acquired.

Recent studies in the auditing literature suggest that auditors with longer tenure are

associated with higher earnings quality (Geiger and Raghunandan, 2002; Gul et al., 2007). Myers et al., (2003) found out that a longer auditor-client relationship is associated with a lower

dispersion of discretionary and current accruals.

Initially the psychologist Hermann Ebbinghaus introduced the learning curve in 1885. Which the psychologist Arthur Bills has described in more detail in 1934 (Brooks et al., 2012). More recent evidence for learning by experience is described in psychology articles (Glaser and Bassok, 1989 and Lapre et al., 2000) which is consistent with the findings in auditing literature. These studies suggest that it takes time for auditors to get client-specific knowledge. For example, during the first year of the audit, the auditor has to deal with questions regarding the non-critical accounting issues and is therefore not able to focus on the more complex and critical accounting issues that may require more attention.

There are also other explanations for the dispersion of discretionary accruals and current accruals, such as the fact that the auditor is less independent as a result of low-balling. The idea behind low-balling is that if an auditor charges lower audit fees to acquire new clients and then expect to recoup these losses in the next years of the engagements (DeAngelo, 1981a). Another explanation could be that clients may switch to another auditor due to poor quality earnings(Gul et al., 2009). High-quality auditors may also drop clients with large unexpected accruals or earnings quality in the first few years of their audit engagements (Gul et al., 2009). This

(11)

11

interpretation is consistent with the auditor switching and opinion shopping literature (e.g. Defond and Subramanyam, 1998; Krishan, 1994; Lennox, 2000)

On the contrary, prior literature, as well as the concerns of regulators and constituents, suggests that a bonding effect has a negative effect on the audit quality. The bonding effect is the risk that due to the bonding between auditors and managers a decrease audit quality in later years will be noted (Brooks et al., 2011).

Several recent studies (Chi and Huang, 2005, Brooks, 2011) found empirical evidence for earnings quality increases in the first years of the audit tenure, and for later deteriorates. Those effects support the view that a learning effect is appearing during the first years of the audit tenure and a bonding effect during the later years. Chi and Huang (2005) found, using Taiwanese data, that during the first five years discretionary accruals decreases and later increases. Brooks (2011) reports that accruals quality increases in early years using US data for both the pre-SOX and the post-SOX eras. They also found that the accruals quality decreases in later years.

2.3 Industry specialization

Prior research has documented that auditor industry specialists provide superior audit quality due to two reasons: 1) they possess in-depth industry knowledge, and hence better ability to provide quality audits; 2) they have incentives to do so due to higher reputation capital (Brooks et al., 2012). Gul et al. (2009) believes that the association between shorter auditor tenure and discretionary accruals is weaker for firms audited by industry specialist. Lim and Tan (2010) found that industry specialists moderate the negative effects of economic bonding on audit quality for long tenure, indicating that the bonding effect is less severe for industry specialists.

DeAngelo (1981b) suggest that auditor specialist has greater concern for loss of

reputation since they have a greater potential loss from audit failures. This is due to the fact that industry specialists invest more in technologies, physical, facilities, personnel and organization control systems that improve the quality of audits in the firms’ local industries.

(12)

12

2.4 Hypotheses

In the previous sections the existing research into this subject was described, I have discussed the motivation for this study and outlined the motivation for this matter. In this paragraph I will discuss and develop the hypotheses.

I have introduced the theory of learning in a previous section. Based on those theories I expect the audit quality to increase in earlier years of the audit tenure. Therefore the first hypotheses will be:

H1 Audit quality increases in earlier years of auditor tenure due to the dominant learning effect.

In a previous section I have discussed the theory that auditor industry specialist having higher audit quality compared to non-specialists. Based on that theory I expect the audit quality for a company whose auditor is an industry specialist is higher, therefore the second hypotheses will be:

H2 Audit quality is higher when audited by the industry specialist.

Since an auditor industry specialist has more experience during the first years of the audit. I expect the audit quality will be higher compared to non-specialist. I expect that industry specialists which have a short tenure audit relationship are associated with a higher amount of audit quality comparing to non industry specialist. Therefore I examine the next hypotheses:

H3 The audit quality of short auditor tenure is higher when audited by an auditor industry specialist.

To examine the risk of lower audit quality due to bonding effects, I will also investigate whether the expected impact of bonding, still exist by long audit tenure of the industry specialization. Seen the results of Lim and Tan (2010) I expect that the bonding effect is less severe for industry specialists. Therefore the next hypotheses is developed:

H4 The audit quality of long auditor tenure is higher when audited by an auditor industry specialist.

2.5 Recap

During this chapter I have discussed the prior research performed on audit quality, learning effect, bonding effect and auditor industry specialization. After that literature review I have developed the hypotheses which are researched in this study. In the next chapter I will have a more detailed look at the research design.

(13)

13

3 Research Design

In this chapter, I explain the research methodology used to test the hypotheses. First, the research sample is clarified. In the last section of this chapter. Secondly, the estimation of audit quality will be explained and the used empirical method is provided.

3.1 Sample

I use a sample of Compustat fundamental Annual files for the years 2002-2012 to collect the data. All U.S. Firms included in the database are included in the sample. There are two reasons for choosing the 2002-2012 time frame. The first is that the implementation of SOX occurred in 2002 and some major accounting scandals occurred in 2001. This master thesis does not want to test the influence of SOX. Secondly, the selection of this period provides eleven observation years which should be provide enough observations to investigate the relation between audit quality and auditor industry specialization and tenure.

The companies are classified by the standard industrial classification codes (SIC), which indicates the type of business of the companies (SEC, 2013). Consistent with Frankel et al. (2002), financial institutions (SIC codes 6000-6999) are not included in the sample because unique procedures are required to estimate discretionary accruals for this type of firms.

Furthermore, the sample will also be restricted to firms with unavailable industry specialization data, and for each two-digit SIC grouping at least eight observations are required. Moreover, it is required that each firm-year observation has the data available to calculate the discretionary accruals metrics and industry specialization metrics. I also deleted observations with auditor code ‘other’ (09). Since the other code is used for all small audit firms, I am not able to calculate the tenure for those observations.

After estimation the abnormal accruals, the first seven years are excluded from the sample since we are not able to determine the tenure for those observations.

3.2 Empirical model

As explained in section 2.1, audit quality can be measured in several ways. In this study I use the modified Jones (1991) discretionary accruals model.

(14)

14

3.2.1 Discretionary accruals model

Since earnings quality is not directly observable a great variety of proxies are used in the approximation of this concept. The proxy for earnings quality used during this research is the amount of discretionary accruals. The level of discretionary accruals is used since it is a ‘direct’ measure of earnings management (Becker et al., 1998). Non-discretionary accruals are the amount of accruals recognized in a specific period in consequence of generally accepted accounting principles. Discretionary accruals contain the part which could be used by the management of the company to manage the earnings.

Prior research used several accrual-based models to determine the amount of earnings management in the financial reporting. Dechow et al. (1995)compare some models to determine the most powerful accrual-model. Based on that research the Jones-model (Jones, 1991) is assess as the most powerful method to estimate the amount of discretionary accruals. Hence, I also use the Jones-model to determine the amount of discretionary accruals.

As outlined above the audit quality can be measured by the amount of discretionary

accruals. More discretionary accruals indicate less audit quality. To estimate discretionary accruals I have to determine the total accruals of a company. This can be measured as follows: Total accruals = Net income -/- operational cash flow (Jones, 1991). When I have determined the amount of total accruals, I can divide them into normal accruals and discretionary accruals by using the modified Jones model. Non-discretionary accruals are estimated using the following model:

TACC = β*a1 (1/TA) + β*a2 ((Δ REV- Δ REC)/TA) + β*a3 (PPE / TA) +

ε

Where:

TACC = Total accruals for sample firm i in year t (scaled by total assets at t-1) TA = Total assets of sample firm i in year t-1

Δ REV = Revenues of sample firm i in year t, minus revenues of sample firm i in year t-1 (scaled by total assets at t-1)

Δ REC = Net receivables of sample firm i in year t, minus net receivables of sample firm i in year t-1 (scaled by total assets at t-1)

PPE = Gross property plant and equipment of sample firm i in year t (scaled by total assets at t-1)

(15)

15

ε = The residual if sample firm i in year t is the proportion of discretionary accruals (positive = income increasing, negative = income decreasing)

All variables are scaled by total assets to control for heteroscedasticity (Jones, 1991). The residual represents the amount of discretionary accruals for firm i in year t. The absolute value of the residual is the estimation of audit quality. A higher value of accrual quality indicates higher audit quality.

3.2.2 Industry specialization

In this study I use five different measures for industry specialization. Since auditor industry specialization is not directly observable, prior work has used several proxies to measure industry specialization. These measures are mostly variants of market share, based on the assumption that industry expertise is built by repetition in similar settings. Following Balsam (2003) a large volume of business in an industry indicates expertise.

In most recent studies a firm’s industry specialization is measured based on its share of clients’ total assets in the two-digit SIC industry group, and industry expertise is assumed when the audit firm’s market share is the highest within the industry group (Krishnan, 2005).

However, market share is subject to several limitations (Gramling et al., 2001). For instance, the audit of a few large clients in an industry could be less advantage over auditing a large

number of smaller clients. To address these shortcomings, I use five proxies to measure industry specialization. Firstly, I use the number of clients as the base. Such a base avoids the bias toward large clients that is implied by using sales as the base (Balsam, 2003). Therefore my first measure identifies an industry specialist as the auditor with the greatest number of clients in the industry.

Secondly in line with Krishan (2005) I assume an industry specialist when the audit firm’s market share is the highest within the industry group. Operationally, I rank auditors based on their percentage of total assets audited in the industry, and the audit firm capturing the largest market share (of total assets) is identified as a specialist in that industry (Gul et al., 2009).

Third: I use as a proxy for industry specialization based on total sales of the client to calculate the market share. I rank auditors based on their percentage of total sales audited in the industry, and the audit firm capturing the largest market share (of total sales) is identified as a specialist in that industry.

(16)

16

Consistent with DeAngelo’s (1981) argument that audit quality increases with audit firm size, the fourth and fifth measure of industry specialization assumes that industry expertise increases with a sufficiently large industry market share has stronger incentives to provide higher audit quality by investing in industry-specific specialization costs (Reichelt and Wang, 2010). Moreover, the fourth and fifth variable defines a industry specialist if in a particular year the auditor has a market share greater than 30% in a two-digit SIC group. Therefore I define all auditors with more than 30% total assets or revenue as a industry specialist.

3.2.3 Auditor tenure

I define auditor tenure as the number of years an auditor is retained by the firm. To test H3 and H4 I divide auditor tenure in three groups. Following prior research (e.g. Johnson et al., 2002) I define auditor tenure as short when the same auditor has audited the financial statements for one to three years. I define the audit tenure as medium when the same auditor has audited the financial statements for 4 years to 6 years. At last, audit tenure more than 7 years is defined as long audit tenure.

3.2.4 Regression model

The multivariate model regresses the absolute value of discretionary accruals on the industry specialization variable and control variables based on prior work (Balsam, 2003, Myers et al., 2003 and Brooks, 2011):

AQ= β0 + β*short_ten + β*long_ten + β*SPEC+ β*ST_SPEC + β*LT_SPEC +

β*SIZE+ β*CFO + β*LEV +β*ROA+

ε

Where:

AQ = Accrual quality, measured as (-1)* absolute value of the residual from the modified Jones model (see above);

Short ten = Short tenure is the number of consecutive years that the firm has retained auditor; 1 if the auditor of firm i has remain the auditor between 1 and 3 years and 0 otherwise

Long ten = Long tenure is the number of consecutive years that the firm has retained auditor; 1 if the auditor of firm i has remain the auditor between 7 years or more and 0 otherwise;

(17)

17

SPEC = 1 if the auditor of firm i is the national-level industry specialist in year t, and 0 otherwise;

ST_SPEC = Short tenure specialist is the number of consecutive years that the firm has been audited by an industry specialist auditor; 1 if the auditor is industry specialist and auditing the firm between 1 and 3 years and 0 otherwise;

LT_SPEC = Short tenure specialist is the number of consecutive years that the firm has been audited by an industry specialist auditor; 1 if the auditor is industry specialist and auditing the firm between 7 and more years and 0 otherwise;

SIZE =The log of the total assets of sample firm i in year t; CFO = Cash flow from operations scaled by total assets year t;

LEV = Total liabilities of sample firm i in year t scaled by total assets year t; ROA =The net income of sample firm i in year t, scaled by the total assets t-1; To control for other determinants that could potentially affect audit quality, I include some control variables based on prior literature. Following Myers et al. (2003), I control for the variables firm size, operating cash flow, total liabilities and ROA.

The control variables for firm size are included since it is likely that accruals quality increases with firm size due to better stability and diversification of activities (Dechow and Dichev, 2002). I control for cash flow from operations since firms with a higher cash flow are more likely to be better performers and have fewer incentives to use earnings management (Frankel et al. 2002;Chen et al., 2008. A high return on assets ratio (ROA) suggest a lower level of discretionary accruals, since the company has less motives to manage the earnings. Prior literature have found that the ROA improves the modified Jones model (Dechow et al,.1995. Kothari et al., 2005).

Leverage is also included since firms with a higher level of debt are associated with the use of more discretionary accruals (Jackson et al., 2008).

(18)

18

3.3 Recap

During this chapter the research design is described. First, I describe the way that the sample is obtained. Secondly the estimation model of discretionary accruals is discussed, the industry specialist variables are outlined and the tenure variables is described. Finally the regression model is discussed. In the next chapter the results are outlined and analyzed.

(19)

19

4 Empirical results and analysis

In this section the final sample is described. Furthermore the various sample distributions are provided. Next the variables are tested for mulitcollinearty and finally the empirical results are provided and analyzed.

4.1 Descriptive statistics

4.1.1 Sample

In the first section of this chapter the sample is described. The sample used in this master thesis is extracted out of the Compustat Annual Files databases. The selection is based on all US companies available in the databases between 2002 and 2012. This resulted in a total of 133.283 unique observations.

The used companies are classified by the standard industrial classification codes (SIC), which indicates the company his type of business (SEC, 2013). Financial institutions (SIC codes 6000-6999) are not included in the sample because unique procedures are required to estimate discretionary accruals for this type of firms. Furthermore firm year observation that did meet the requirements to calculate the earnings management metrics were deleted. And all observations with auditor code ‘other’ are also deleted from the sample. Furthermore, all the firm year

observations that did not meet the criteria to calculate the industry specialization variables and all observations with an auditor variable 0 (unaudited) are deleted from the sample. Finally, the years 2002-2007 are deleted from the sample, since the tenure of the auditor cannot be determined. This resulted in a final sample with 18.236 observations as presented in table 1:

Table 1: Final Sample

Calculation of final sample

Observations Firm year observations 2002-2012 Compustat Annual files 133.283 Deleting missing data industry specialization and ‘unaudited’ data 314

Deleting missing data to calculate AEM variable 60.275

Deleting missing data, financial institutions and industry year with less than

8 observations and auditor other 33.915

Observations used to calculate AEM variable 38.779

Deleting year observations 2002-2007 20.543

(20)

20

Table 2 shows the observations per year. The distribution over the year indicates that the observations are equally divided over the research years. The year with the least firm

observations is 2012 with 3.540 observations that meet all the criteria. The year 2008 has the most observations with 3.819 observations.

Table 2: sample distribution

Distributions of observations per year

Fiscal year Freq. Percent

2008 3.819 20,94 2009 3.698 20,28 2010 3.617 19,83 2011 3.562 19,53 2012 3.540 19,41 Total 18.236 100

The sample contains firms from eight different industries. The Manufacturing industry has the largest amount of observations (8.212) while the Agriculture, Forestry and Fishing industry has the smallest amount of observations (50).

Table 3: sample distribution

Distribution of observations per industry

Industry Freq. Percent

Agriculture, Forestry, & Fishing 50 0,27 Construction 203 1,11 Manufactering 8.212 45,03

Mining 2.540 13,93

Retail Trade 1.256 6,89 Service Industries 3.235 17,75 Transportation & Public Utilities 2.120 11,63 Wholesale Trade 620 3,40

(21)

21

Figure 2: Distribution of observations per industry

A more detailed look at the industry distribution gives that the sample contains firms from 54 different industries based on two-digit SIC Codes. Industry 51 (Building Materials & Gardening Supplies) has the smallest amount of observations (26) and industry 73 (Business services) has the most observations (2.073) in the sample. In appendix 1, an overview of the industry classifications is provided in a frequency distribution.

The most observations for the different auditors, as showed in table 4, are so-called Big4 auditors (Deliotte, EY, KPMG and PWC) 84%. The auditor with the most observations is EY (4.546) followed by PWC (3.922). After the Big4 the fifth and sixth largest observations are from Grant Thorton (1.228) and BDO (888), the observations from other auditors are relative low, 878 (4.6%) observations in total.

Agriculture, Forestry, & Fishing 0% Construction 1% Manufactering 45% Mining 14% Retail Trade 7% Service Industries 18%

Transportation & Public Utilities

12% Wholesale Trade

3%

(22)

22

Table 4: Descriptive statistics

Auditor distribution

Auditor Name Auditor Code Freq. Percent

EY 4 4.546 24,98 Deliotte 5 3.364 18,49 KPMG 6 3.410 18,74 PWC 7 3.922 21,56 Touche Ross 8 2 0,01 BDO 11 888 4,88

Baird, Kurtz and Dobson 12 6 0,03

Cherry, Bekaert and Holland 13 32 0,18

Crowe Chizek 16 96 0,53

Grant Thornton 17 1.228 6,75

J H Cohn 18 103 0,57

McGladrey and Pullen 21 312 1,71

Moore Stephens 22 60 0,3

Moss Adams 23 122 0,67

Pannell Kerr Foster 24 46 0,2

Plante & Moran 25 35 0,12

Richard A. Eisner 26 64 0,29

Total 18.236 100

Figure 3 shows the percentage of industry specialist for the several measurements. The highest percentage (29,7%) is for auditor industry specialist based on the highest total assets. The lowest percentage (18,8) is for auditor industry specialist which are based on auditing total assets greater than 30%.

(23)

23

In table 5 the descriptive statistics of all variables are outlined. All variables used in the calculation for the various earnings management metrics, industry specialization and tenure. Those variables are windsorized to reduce the effect of possibly spurious outliers. All data below the first percentile are set to the first percentile and all data above the percentile are set to the 99th percentile. Table 5 shows that the mean tenure is 6.56 year.

Table 5: Descriptive Statistics

Frequency, count, mean, sd of variables

count mean sd min max

AEM 18236 0.04 0.31 -8.2 7.5

Spec. based on Number of

clients 18236 0.28 0.45 0.0 1.0

Spec. based on highest TA 18236 0.30 0.46 0.0 1.0

Spec. based on TA >30% 18236 0.19 0.39 0.0 1.0

Spec. based on highest revt 18236 0.24 0.43 0.0 1.0

Spec. based on revt >30% 18236 0.27 0.44 0.0 1.0

Tenure 18236 6.56 2.87 1.0 11.0

Log tenure 18236 1.74 0.61 0.0 2.4

Short tenure 18236 0.19 0.40 0.0 1.0

Long tenure 18236 0.58 0.49 0.0 1.0

Size Total Assets 18236 6.19 2.15 -4.0 12.7

Leverage 18236 1.02 57.88 0.0 7,804.0

Return on Assets 18236 -0.14 9.32 -1,250.0 24.9

Cash flow from operations 18236 0.03 0.35 -6.2 0.6

N 18236 5,0 10,0 15,0 20,0 25,0 30,0 35,0 1 % Figure 3 % of industry specialist

Spec. based on Number of clients Spec. based on highest TA Spec. based on TA >30%

(24)

24

4.2 Correlation matrix

To test the relationship between each variable a test for multicollinearity is performed. In the correlation matrix outlined in table 6 the Pearson value of correlation can vary between -1 and 1. This will either indicate a negative or positive correlation of the two variables. If the Pearson coefficient is zero there is no correlation between the two variables indicated. As a rule of thumb used in common literature (Field, 2009), Pearson coefficients higher than 0.7 or lower than -0.7 will have an effect of the reliability of the model. Table 6 shows that auditor industry specialist based on number of clients is positively significant correlated (0.917) with an auditor industry specialist based on auditing the highest amount of total assets. Auditor industry specialist based on number of auditing more than 30% of total revenue is positively significant correlated (0.806) to auditor industry specialist based on highest amount of revenue audited in a two-digit SIC industry code. All variables used for long tenure specialist are significant positively correlated to the specialist variable.

In line with the positively significant correlated specialist variables described above the tenure, short- and long tenure, are positively significant correlated. Beside that I noted a significant positive correlation (0.702) for the variable short tenure of an auditor industry specialist auditing 30% of total assets and the auditor industry specialist based on number of clients.

Tenure is significant positively correlated to the tenure (0.952), while long tenure is significant positively correlation with tenure (0.856) as log of tenure (0.788). Furthermore is short tenure significant negatively correlation to tenure (-0.752) and the log of tenure (-0.844). Which is conform our expectations.

Since the correlated industry specialist variables are not in the same regression, there is no need to eliminate the variables from the regression. In the next section the industry specialists regressions will be discussed.

(25)

Table 6: Pearson Correlation Matrix 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 1. AEM 1.000 2.Return on Assets -0.0966* 1.000 3. Leverage 0.0583* -0.9955* 1.000 4. Total accruals 0.0080 -0.0013 0.0013 1.000 5. Cash flow from operations

-0.348* 0.201* -0.155* -0.0232* 1.000 6. Delta Revenu

0.0015 0.0411* -0.0346* -0.0019 0.1137* 1.000 7. Log total assets

-0.1128* 0.0533* -0.0341* 0.0018 0.3749* 0.0508* 1.000 8. Spec. based on Number of

clients -0.0195* 0.0061 -0.0057 0.0011 0.0023 -0.0086 0.0791* 1.000 9. Spec. based on highest TA

-0.0209* 0.0067 -0.0059 -0.0000 0.0097 -0.0046 0.1051* 0.9169* 1.000 10. Spec. based on TA >30%

-0.0129 0.0043 -0.0041 0.0036 -0.0266* -0.0148* 0.1057* 0.6468* 0.6371* 1.000 11. Spec. based on highest

revt -0.0142 0.0070 -0.0048 -0.0010 0.0496* -0.0047 0.1671* 0.2626* 0.3049* 0.2764* 1.000 12. Spec. based on revt >30%

-0.0151* 0.0070 -0.0051 -0.0006 0.0383* 0.0012 0.1971* 0.2118* 0.2496* 0.3582* 0.8063* 1.000 13. Tenure -0.0453* 0.0216* -0.0131 0.0044 0.1294* -0.0193* 0.3445* 0.0653* 0.0810* 0.0871* 0.0958* 0.1091* 1.000 14. Log Tenure -0.0516* 0.0237* -0.0142 0.0039 0.1353* -0.0248* 0.3247* 0.0718* 0.0856* 0.0820* 0.0891* 0.0980* 0.9518* 1.000 15. Long Tenure -0.0381* 0.0169* -0.0101 -0.0000 0.1159* -0.0267* 0.3185* 0.0597* 0.0727* 0.0769* 0.1003* 0.1109* 0.8561* 0.7876* 1.000 16. Short Tenure 0.0428* -0.0240* 0.0161* -0.0021 -0.1117* 0.0299* -0.2493* -0.0609* -0.0674* -0.0601* -0.0576* -0.0631* -0.7522* -0.8443* -0.5780* 1.000 17. Tenure Spec. based on

Number of clients long

-0.0220* 0.0066 -0.0041 0.0003 0.0360* -0.0197* 0.1581* 0.7393* 0.6820* 0.5047* 0.2219* 0.1869* 0.3394* 0.3119* 0.3955* -0.2286* 1.000 18. Tenure Spec. based on

highest TA long

-0.0233* 0.0070 -0.0042 -0.0006 0.0410* -0.0164* 0.1798* 0.6716* 0.7422* 0.4912* 0.2541* 0.2169* 0.3568* 0.3260* 0.4101* -0.2370* 0.9250* 1.000 19. Tenure Spec. based on

TA >30% long

-0.0187* 0.0053 -0.0031 0.0015 0.0150* -0.0211* 0.1612* 0.4963* 0.4900* 0.7817* 0.2257* 0.3003* 0.2888* 0.2600* 0.3196* -0.1848* 0.6889* 0.6761* 1.000 20. Tenure Spec. based on

highest revt long

-0.0195* 0.0071 -0.0037 -0.0022 0.0659* -0.0145 0.2144* 0.2022* 0.2393* 0.2187* 0.7765* 0.6293* 0.3296* 0.3000* 0.3750* -0.2168* 0.3603* 0.4014* 0.3450* 1.000 21. Tenure Spec. based on

revt >30% long

-0.0219* 0.0074 -0.0038 -0.0019 0.0622* -0.0103 0.2446* 0.1637* 0.1981* 0.2881* 0.6228* 0.7747* 0.3541* 0.3201* 0.3960* -0.2289* 0.3217* 0.3602* 0.4321* 0.8267* 1.000 22. Tenure Spec. based on

Number of clients short

0.0045 -0.0003 -0.0021 -0.0004 -0.0395* 0.0156* -0.0677* 0.3422* 0.3098* 0.2042* 0.0831* 0.0661* -0.3217* -0.3496* -0.2532* 0.4381* -0.1002* -0.1038* -0.0809* -0.0950* -0.1003* 1.000 23. Tenure Spec. based on

highest TA short

0.0044 -0.0001 -0.0022 -0.0005 -0.0368* 0.0152* -0.0615* 0.3123* 0.3363* 0.2019* 0.0981* 0.0794* -0.3248* -0.3508* -0.2570* 0.4446* -0.1016* -0.1054* -0.0821* -0.0964* -0.1018* 0.9381* 1.000 24. Tenure Spec. based on

TA >30% short

0.0100 -0.0009 -0.0016 -0.0003 -0.0511* 0.0018 -0.0401* 0.2317* 0.2274* 0.3481* 0.1127* 0.1282* 0.2513* - -0.2757* -0.1969* 0.3406* -0.0779* -0.0807* -0.0629* -0.0738* -0.0780* 0.7021* 0.6998* 1.000 25. Tenure Spec. based on

highest revt short

0.0068 0.0005 -0.0020 -0.0008 -0.0135 0.0110 -0.0300* 0.0870* 0.1003* 0.1003* 0.3483* 0.2834* -0.2975* -0.3264* -0.2327* 0.4026* -0.0920* -0.0954* -0.0744* -0.0873* -0.0922* 0.4207* 0.4581* 0.4356* 1.000 26. Tenure Spec. based on

revt >30% short

(26)

4.3 Test of industry specialist

In this section of chapter 4 the regression model as outlined in 3.2.4 is tested. Furtermore I will discuss the main findings.

4.3.1 Industry specialist and tenure model

In this section the association between the auditor industry specialist metrics is tested in the regression model. In table 10 the estimated coefficients are shown. The table gives an overview of the various variables and if the results from the regression models are statistically significant. In the next section of this chapter the outcome is outlined.

4.3.1.1 AEM – Auditor industry specialist and tenure

The model is tested where AEM is the dependent variable is tested. The robustness regression coefficients are estimated using the following model:

AQ= β0 + β*short_ten + β*long_ten + β*SPEC+ β*ST_SPEC + β*LT_SPEC +

(27)

27

Table 7: Regression models

(1) (2) (3) (4) (5) Spec. based on Number of clients Spec. based on highest TA Spec. based on TA >30% Spec. based on highest

revt Spec. based on revt >30%

VARIABLES AEM AEM AEM AEM AEM

Constant 0.012*** 0.012*** 0.012*** 0.012*** 0.012*** (0.000) (0.000) (0.000) (0.000) (0.000) Short tenure 0.001** 0.001** 0.001*** 0.001** 0.001** (0.000) (0.000) (0.000) (0.000) (0.000) Long_tenure 0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) Spec. based on Number of

clients 0.000

(0.000) Short tenure specialist based

on number of clients 0.000 (0.001) Long tenure specialist based

on number of clients -0.001* (0.000)

Spec based on highest TA 0.000

on highest TA 0.000

(0.000) Long tenure specialist based

on highest TA -0.001**

(0.000)

Spec based on TA >30% -0.001

on TA >30% 0.000

on TA >30% -0.001**

(0.000)

Spec based on highest revt -0.000

on highest revt 0.001

on highest revt -0.000

(0.000)

Spec based on revt >30% -0.000

on revt >30% 0.001

on revt >30% -0.001

(0.000) Size -0.001*** -0.001*** -0.001*** -0.001*** -0.001***

(0.000) (0.000) (0.000) (0.000) (0.000) Cash flow from operations -0.016*** -0.016*** -0.017*** -0.016*** -0.016***

(0.000) (0.000) (0.000) (0.000) (0.000) Leverage -0.000*** -0.000*** -0.000*** -0.000*** -0.000*** (0.000) (0.000) (0.000) (0.000) (0.000) Return on Assets 0.012*** 0.012*** 0.012*** 0.012*** 0.012*** (0.000) (0.000) (0.000) (0.000) (0.000) Observations 18,226 18,226 18,226 18,226 18,226 R-squared 0.461 0.461 0.470 0.461 0.467

Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1

(28)

28

The regression is tested whereby the auditor industry specialist variables are determined based on the variables outlined in section 3.2.2. Table 7 shows that the R-squared of this model is 46,1% which indicate that the explanatory power of the model is quite good. Furthermore I found that short tenure is significant positive association with accrual based earnings, which indicates that short tenure has negative effects on audit quality. We found for three industry specialist variables; auditor industry specialist based on number of clients in a two-digit SIC industry code, auditors industry specialist based on auditing the highest amount of total assets in a digit SIC industry code and the industry specialist auditing 30% of the total assets in a two-digit SIC industry code, that the long tenure specialist has a significant negative effect on the accrual based earnings management which indicates a higher audit quality. In addition the following control variables are statistically significant:, Size, Cash flow from operations, Leverage and Return on Assets.

The variable cash flow from operations is significant in a negative direction which is consistent with prior research (Balsam, 2003). The variable ROA is significant in a positive direction. This is consistent with the results from Kothari, Leone, and Walsey (2005). In their research they argue that an indication of abnormal discretionary accruals is higher in firms with unusual performance. The variable leverage is significant in a negative direction which is consistent with the results of Balsam et al. (2003). The variable Size is significant in a negative direction which is also in line with the results of Balsam et al. (2003).

4.4 Analyse

In this section an analysis of the results presented in chapter 4.3 of this master thesis will be given. The outcomes of the regression models provide a source of information to determine if the in chapter 2 developed hypotheses cannot be rejected.

Firstly the results shows that short tenure can be associated with higher accrual based earnings management and therefore lower audit quality. Therefore it can be stated that short tenure indicates a lower audit quality. From these observations it can be stated that hypothesis 1, audit quality increases in earlier years of auditor tenure due to the dominant learning effect, cannot be rejected.

The results does not indicates that industry specialization could be associated with higher audit quality. I find a negative effect on accrual based earnings management for three auditor industry specialist variables, which could indicate a positive effect on audit quality for industry

(29)

29

auditor specialist based on auditing more than 30% of a two-digit SIC industry code, auditing the most revenue in a two-digit SIC industry and auditing more than 30% of a two-digit SIC industry code. Seen the fact that those effect where not significant no supporting evidence for hypothesis 2 is found, therefore hypothesis 2, auditor industry specialist can be associated with higher audit quality, can be rejected.

For the variable short tenure industry specialist I find positive effects on accrual based earnings management, indicating lower audit quality. But these result were not significant, seen those result hypothesis 3, the audit quality of short auditor tenure is higher when audited by an

auditor industry specialist, can be rejected.

Hypothesis 4 stated: The audit quality of long auditor tenure is higher when audited by an

auditor industry specialist. I find for three regression models, specialist based on auditing the highest number of clients in a two-digit SIC industry code, highest amount of total assets and auditing more than 30% of total assets, a significant negative effect on accrual based earnings management which indicates a higher audit quality. I also find a negative effect on accrual based earnings management for the last two auditor industry specialist variables but these results were not significant. Seen the majority of auditors industry specialist indicates a significant effect hypothesis 4 cannot be rejected.

4.5 Recap

In this chapter the sample statistics were described and tested for multicollinearity. After these tests the audit quality metric was estimated using various auditor industry specialist variables for every two-digit SIC industry code with more than 8 observations and a defined auditor. The auditor industry model shows the results of the regressions which are analyzed in section 4.4. In next chapter the conclusions will be drawn.

(30)

30

5 Conclusion

This section will start with a conclusion about the relationship between auditor tenure and auditor industry specialization on audit quality. The chapter also contains a section describing the limitations of this research, what the contribution to academic literature and society is and suggestions for future research.

5.1 Conclusions

In this section conclusions are drawn from the analysis of the results outlined in section 4.4. The analysis provide evidence that short audit tenure and audit quality are associated. short audit tenure indicates a lower audit quality since it has a positive impact on accrual earnings

management. I cannot conclude that auditor industry specialization has a significant effect on audit quality. Furthermore I cannot conclude that short tenure when audited by an auditor industry specialist is associated with higher audit quality. I can conclude that long audit tenure audit by an industry specialist can be associated with higher audit quality. A possible explanation of these results might be that auditors with industry expertise in the client’s business are more likely to report irregularities and misrepresentations and provide higher audit quality seen the fact that there is a greater loss of reputation when the audit quality is too low.

So, this study provides evidence, as indicate in early papers (i.e. Beck and Wu, 2006), that short tenure could association with lower audit quality. The results of this study does not provide evidence that auditor industry specialization is associated with higher audit quality.

Besides that, this master thesis provides evidence that long tenure audit relationship with an auditor industry specialist, based on three industry specialist variables, positively affects audit quality as indicated earlier by Lim and Tan (2010). Therefore it can be stated that a negative effect on bonding occurs due to auditors industry specialization. This master thesis does not provide evidence if auditor industry specialist audit a client with a short audit tenure, it is associated with higher audit quality. Therefore no conclusion can be drawn about the possible learning effect and industry specialization.

5.2 Limitations

The limitations of this research are primarily caused by the time and scope of this master thesis. The first limitation is the sample selection. The final sample contains only firm years between

(31)

31

2002-2012. Because this time span is relative small for tenure measurement this could introduce a bias in the results of this master thesis. Especially, seen the factors of tenure can only be measured for the last five years of the sample.

The second limitation is the years included in the sample selection. The first years in the sample are after the implementation of SOX and the great accounting scandals which could lead to conservative accounting during these years. Beside that some years in the sample are

influenced by the global economic crisis.

Besides that, I only used discretionary accruals as a measure of audit quality. It could be possible that other ways to measure audit quality, such as propensity to issue a going-concern audit-opinion for distressed companies or the earnings response coefficient, would have led to other results.

I only use data from the United States. Therefore results have to be interpreted with caution when generalized to other countries, since there are notable differences in the institutional backgrounds and the power of enforcement worldwide.

5.3 Contributions to academic literature and society

In this master thesis the relationship between auditor industry specialization, audit tenure and audit quality is investigated. My thesis does shows a lesser extent of bonding effect by auditor industry specialist. My thesis does not provide evidence for a lesser extent of learning effect by auditor industry specialists. The result from this master thesis can help investors and auditors to assess the quality of financial information when an auditor industry specialist is engaged.

5.4 Suggestions for future research

As mentioned in section 5.3 the timeframe is relative small and in further research this could be extended. Furthermore this research can be reperformed in another country. The third

suggestion is the use of other auditor industry specialist variables such as auditor industry specialist based on total amount of audit fees in a two-digit SIC industry group, or a distinguish between national-office specialist or local-office specialist, or following Mayhew and Wilkins (2002), defining an auditor industry specialist if they are the largest supplier in the industry and the difference between the first and second supplier in the industry is at least 10 percent. At last this subject could further researched by using other audit quality measures.

(32)

32

Literature

Balsam, S., J. Krishan and S. Joon Yang. 2003. Auditor industry specialization and earnings quality. Auditing: A journal of practice & theory. Vol. 22 No. 2 PP.71-97

Beasley, M. S., & Petroni, K. R. (2001). Board independence and audit-firm type. Auditing: A Journal of Practice & Theory, 20(1), 97-114.

Beck, P. J., & Wu, M. G. (2006). Learning by Doing and Audit Quality*. Contemporary Accounting Research, 23(1), 1-30.

Becker, C. L., DeFond, M. L., Jiambalvo, J., & Subramanyam, K. R. (1998). The Effect of Audit Quality on Earnings Management*. Contemporary accounting research, 15(1), 1-24.

Brooks, L. Z., Cheng, C. S., Johnston, J. A., & Reichelt, K. J. (2011). When Does Audit Quality Start to Decline in Firm Audit Tenure?-An International Analysis. An International Analysis (September 25, 2011).

Brooks, L., Cheng, A., & Reichelt, K. (2012). Audit Firm Tenure and Audit Quality: Evidence from US Firms. Kenneth J., Audit Firm Tenure and Audit Quality: Evidence from US Firms (April 10, 2012).

Browning, E.S., and J. Weil. 2002. Accounting woes roil stock markets as nervous investors stampede exits. Wall Street journal (January 29)

Carcello, J. V., & Nagy, A. L. (2002). Auditor industry specialization and fraudulent financial reporting. In Symposium on Auditing Problems.

Chen, C. Y., Lin, C. J., & Lin, Y. C. (2008). Audit Partner Tenure, Audit Firm Tenure, and Discretionary Accruals: Does Long Auditor Tenure Impair Earnings Quality?*. Contemporary Accounting Research, 25(2), 415-445.

Chen, J. T., & Manes, R. P. (1985). Distinguishing the two forms of the constant percentage learning curve model. Contemporary Accounting Research, 1(2), 242-252.

Chi, W., & Huang, H. (2005). Discretionary accruals, audit-firm tenure and audit-partner tenure: Empirical evidence from Taiwan. Journal of Contemporary Accounting & Economics, 1(1),

65-92.

Craswell, A. T., Francis, J. R., & Taylor, S. L. (1995). Auditor brand name reputations and industry specializations. Journal of accounting and economics, 20(3), 297-322.

(33)

33

DeAngelo, L. E. (1981a). Auditor independence,‘low balling’, and disclosure regulation.

Journal of accounting and Economics, 3(2), 113-127.

DeAngelo, L. E. (1981b). Auditor size and audit quality. Journal of accounting and economics, 3(3), 183-199.

Dechow, P. M., & Dichev, I. D. (2002). The quality of accruals and earnings: The role of accrual estimation errors. The accounting review, 77(s-1), 35-59.

Dechow, P. M., R. G. Sloan, and A. P. Sweeney. (1995). Detecting earnings management. The Accounting Review 70: 193-225.

DeFond, M. L., & Jiambalvo, J. (1994). Debt covenant violation and manipulation of accruals. Journal of accounting and economics, 17(1), 145-176.

DeFond, M. L., & Subramanyam, K. R. (1998). Auditor changes and discretionary accruals.

Journal of Accounting and Economics, 25(1), 35-67.

Defond, M.L., and J.R. Francis. (2005). Audit research after Sabanes-Oxley. Auditing: A Journal of Practice & Theory 24: 5-30.

European Commission (2010). Green paper, Audit Policy: Lessons from the Crisis, Belgium, Brussels.

Field, A. (2009). Discovery statistics using SPSS. Saga publications ltd.

Frankel, R.M., M.F. Johnson and K.K. Nelson (2002). The Relation between auditors' fees for non audit services and earnings management. The Accounting Review, 77, (supplement 2002),

pp. 71-105

Geiger, M. A., & Raghunandan, K. (2002). Auditor tenure and audit reporting failures. Auditing: A Journal of Practice & Theory, 21(1), 67-78.

Glaser, R., & Bassok, M. (1989). Learning theory and the study of instruction (No.

UPITT/LRDC/ONR/KBC-12). PITTSBURGH UNIV PA LEARNING RESEARCH AND DEVELOPMENT CENTER.

Gramling, A. A., & Stone, D. N. (2001). Audit firm industry expertise: A review and synthesis of the archival literature. Journal of Accounting Literature, 20, 1-29.

Gul, F. A., Fung, S. Y. K., & Jaggi, B. (2009). Earnings quality: Some evidence on the role of auditor tenure and auditors’ industry expertise. Journal of Accounting and Economics, 47(3), 265-287.

(34)

34

Gul, F. A., Jaggi, B. L., & Krishnan, G. V. (2007). Auditor independence: evidence on the joint effects of auditor tenure and nonaudit fees. Auditing: A Journal of Practice & Theory, 26(2),

117-142.

Jackson, A. B., Moldrich, M., & Roebuck, P. (2008). Mandatory audit firm rotation and audit quality. Managerial Auditing Journal, 23(5), 420-437.

Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360.

Johnson, V. E., Khurana, I. K., & Reynolds, J. K. (2002). Audit‐Firm Tenure and the Quality of Financial Reports*. Contemporary accounting research, 19(4), 637-660.

Jones, J. (1991). Earnings management during import relief investigations. Journal of Accounting Research, 29, (2), pp. 193-228.

King, R. R., & Schwartz, R. (1999). Legal Penalties and Audit Quality: An Experimental Investigation*. Contemporary Accounting Research, 16(4), 685-710.

Kothari, S. P., Leone, A. J., & Wasley, C. E. (2005). Performance matched discretionary accrual measures. Journal of accounting and economics, 39(1), 163-197.

Krishnan, J. (1994). Auditor switching and conservatism. Accounting Review, 200-215.

Lapré, M. A., Mukherjee, A. S., & Van Wassenhove, L. N. (2000). Behind the learning curve: Linking learning activities to waste reduction. Management Science, 46(5), 597-611.

Lennox, C. (2000). Do companies successfully engage in opinion-shopping? Evidence from the UK. Journal of Accounting and Economics, 29(3), 321-337.

LIM, C. Y., & TAN, H. T. (2010). Does Auditor Tenure Improve Audit Quality? Moderating Effects of Industry Specialization and Fee Dependence*. Contemporary Accounting Research, 27(3), 923-957.

Low, K. Y. (2004). The effects of industry specialization on audit risk assessments and audit-planning decisions. The Accounting Review, 79(1), 201-219.

Mascarenhas, D., Cahan, S. F., & Naiker, V. (2010). The effect of audit specialists on the informativeness of discretionary accruals. Journal of Accounting, Auditing & Finance, 25(1), 53-84.

(35)

35

Mayhew, B. W., & Wilkins, M. S. (2003). Audit firm industry specialization as a

differentiation strategy: Evidence from fees charged to firms going public.Auditing: A Journal of Practice & Theory, 22(2), 33-52.

Morgan, J., & Stocken, P. (1998). The effects of business risk on audit pricing. Review of Accounting Studies, 3(4), 365-385.

Myers, J. N., Myers, L. A., & Omer, T. C. (2003). Exploring the term of the auditor-client relationship and the quality of earnings: A case for mandatory auditor rotation?. The Accounting Review, 78(3), 779-799.

Reichelt, K. J., & Wang, D. (2010). National and office‐specific measures of auditor industry expertise and effects on audit quality. Journal of Accounting Research, 48(3), 647-686.

Reynolds, J. K., & Francis, J. R. (2000). Does size matter? The influence of large clients on office-level auditor reporting decisions. Journal of Accounting and Economics, 30(3), 375-400.

SEC (2013). Division of Corporation Finance: Standard Industrial Classification (SIC) Code List (http://www.sec.gov/info/edgar/siccodes.htm), 15 May.

Simunic, D. A. (1984). Auditing, consulting, and auditor independence. Journal of accounting research, 22(2), 679-702.

Solomon, I., Shields, M. D., & Whittington, O. R. (1999). What do industry-specialist auditors know?. Journal of Accounting Research, 37(1), 191-208.

Yelle, L. E. (1979). The learning curve: Historical review and comprehensive survey. Decision Sciences, 10(2), 302-328.

(36)

36

Appendix 1 Distribution per two-digit SIC industry

Table I

Distribution per two-digit SIC industry

Industry

Two-digit SIC

codes Freq. Percent

Agricultural Production - Crops 1 50 0,27

Metal, Mining 10 972 5,33

Coal Mining 12 87 0,48

Oil & Gas Extraction 13 1.394 7,64

Nonmetallic Minerals, Except Fuels 14 87 0,48

General Building Contractors 15 70 0,38

Heavy Construction, Except Building 16 93 0,51

Special Trade Contractors 17 40 0,22

Food & Kindred Products 20 437 2,4

Textile Mill Products 22 57 0,31

Apparel & Other Textile Products 23 138 0,76

Lumber & Wood Products 24 112 0,61

Furniture & Fixtures 25 115 0,63

Paper & Allied Products 26 193 1,06

Printing & Publishing 27 183 1

Chemical & Allied Products 28 1.937 10,62

Petroleum & Coal Products 29 123 0,67

Rubber & Miscellaneous Plastics Products 30 150 0,82

Leather & Leather Products 31 63 0,35

Stone, Clay, & Glass Products 32 88 0,48

Primary Metal Industries 33 235 1,29

Fabricated Metal Products 34 235 1,29

Industrial Machinery & Equipment 35 988 5,42

Electronic & Other Electric Equipment 36 1.521 8,34

Transportation Equipment 37 448 2,46

Instruments & Related Products 38 1.042 5,71

Miscellaneous Manufacturing Industries 39 147 0,81

Railroad Transportation 40 40 0,22

Trucking & Warehousing 42 136 0,75

Water Transportation 44 248 1,36

Transportation by Air 45 159 0,87

Pipelines, Except Natural Gas 46 52 0,29

Transportation Services 47 70 0,38

Communications 48 642 3,52

Electric, Gas, & Sanitary Services 49 773 4,24

Wholesale Trade - Durable Goods 50 374 2,05

Wholesale Trade - Nondurable Goods 51 246 1,35

Building Materials & Gardening Supplies 52 26 0,14

General Merchandise Stores 53 116 0,64

Food Stores 54 106 0,58

Automative Dealers & Service Stations 55 119 0,65

Apparel & Accessory Stores 56 225 1,23

Furniture & Homefurnishings Stores 57 82 0,45

Eating & Drinking Places 58 260 1,43

(37)

37

Hotels & Other Lodging Places 70 54 0,3

Personal Services 72 67 0,37

Business Services 73 2.073 11,37

Auto Repair, Services, & Parking 75 49 0,27

Motion Pictures 78 75 0,41

Amusement & Recreation Services 79 152 0,83

Health Services 80 331 1,82

Educational Services 82 97 0,53

Engineering & Management Services 87 337 1,85

The effect of industry specialization & audit tenure on audit quality