The Impact of Local Competition on Audit Fees and Audit Quality.

(1)

The Impact of Local Competition on Audit Fees and Audit Quality.

Name: Jack Blaauw Student number: 13317938 Word Count:

Submitted: 20/06/21

MSc Accountancy and Control

Faculty of Economics and Business, University of Amsterdam

(2)

Statement of Originality

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

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

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

(3)

Abstract

This study examines whether local market competition impacts audit pricing and audit quality and its differential effect on Big-4 and Non-Big 4 auditors. I examine this issue using a large sample of U.S. audit clients covering 47 metropolitan statistical areas (MSAs) spanning from 2011–2019. The study is motivated by the concern that high levels of concentration has resulted in lower levels of competition, creating impediments on the audit practice through higher audit fees and lower levels of service. However, a contrary argument is that high levels of concentration creates efficiencies and betterments to the practice, due to the attainment of economies of scale, allowing for lower audit fees and improved quality. I find that local market competition is insignificantly associated with lower audit fees, to which has no differential effect between NB4 and Big-4 auditors, inconsistent with the concerns of regulators. I also find that local market competition has no impact on the audit quality provided, to which no differentiated effect exists between NB4 and Big-4 auditors. I interpret my findings as evidence that even within highly concentrated audit markets, audit firms still compete amongst one another to attain the services of clients. My results are relevant to the ongoing debate regarding the consequences of increased concentration within the U.S. audit market (GAO 2003, 2008).

(4)

Regulators are concerned that greater market concentration has led to non-competitive behaviour by Big-4 accounting firms, resulting in impediments on the audit practice (European Commission, 2010; Financial Reporting Council [FRC], 2018; Government Accountability Office [GAO], 2003, 2018). The concern of regulators, stems from the economies of scale that has been achieved by Big-4 accounting firms, resulting in excessive market power (e.g., Danos

& Eichenseher, 1982; Dopuch & Simunic, 1980; Pound & Francis, 1981; Sirois & Simunic, 2011). It is argued that high market concentration necessitates the audit practice, allowing for higher quality financial reporting. However, as mentioned above, regulators argue that increased market concentration has reduced audit market competition, impeding the practice.

The motivation behind this paper is to address this issue empirically, by examining the impact local market competition¹ has on auditor pricing mechanisms and service of quality provided.

This will be of use to regulators, as high audit market concentration has been labelled a “deep seated problem” (GAO, 2008), and there is no clear evidence to support this claim.

Various papers, have attempted to measure the impact of audit competition on audit fees and quality, assuming that audit market concentration measures market competition (e.g., Pearson

& Trompeter, 1994; Numan & Willekens, 2012). However, industrial organisation theory, argues that market concentration is a static measure of market competition (Carlton & Perloff, 1994). The effect of audit competition on audit fees has been indirectly measured (Maher, Tiessen, Colson, & Broman, 1992; Sanders, Allen, & Korte, 1995; Ghosh & Lustergarter, 2006) and infers that in the presence of audit market competition, audit fees decline noticeably, however, few papers have attempted to use a direct measure of market competition, in which this paper attempt to do.

A similar audit competition measure is employed, following Caves & Porter (1978), Buijink, Maijoor, & Meuwissen, (1998) and Van Raak, Peek, Meuwissen & Schelleman, (2020), in which the absolute value of year over percentage point changes in market concentration are measured, representing market competition.

1 Local market competition refers to the industry, county specific level of competition.

(6)

The sample consists of US audit engagements spanning across 2011 to 2019, at the metropolitan statistical area² level (hereby referred to as MSA). Measuring at the MSA³ level, is shown to be the most effective way to measure presence of competition for multiple reasons, as indicated across literature (e.g. Newton, Wang, & Wilkins, 2013; Numan et al., 2012;

Francis, Stokes, & Anderson, 1999; Reynolds & Francis, 2001). First, measuring MSA allows for greater variation in competition compared to when competition is measured at the national level. Second, decisions regarding audit opinions and pricing are decided at the local office level (Francis et al. 1999; Reynolds et al., 2001). Third, auditors generally only audit offices within close proximity (Choi, Kim, & Zang., 2010). Fourth, auditor expertise and audit quality has been shown to vary across offices within the same firm (Choi et al., 2010; Francis, Michas,

& Seavey, 2013). Furthermore, I measure local competition within industries (market) to control for demand of services required by clients.

Restatements are used as a proxy for audit quality as restatements measure a failure in the primary responsibility of the auditor. An auditor’s responsibility requires them to deliver financial statements free of material misstatement or error, therefore, restatements indicate that auditors have failed to act according to their responsibilities (Newton et al., 2013; Palmrose &

Scholz, 2004). The accounting literature often employs restatements to measure the audit quality possessed by clients (e.g., Kinney, Palmrose, & Scholz, 2004; Schmidt, 2012; Schmidt and Wilkins, 2013; Newton et al., 2013). It is important to acknowledge that no single measure of audit quality is without measurement error. Nonetheless, using restatements as a measure of audit quality allows appropriate conclusions to be drawn, however, results should be approached with caveat.

Contrary to regulators concerns over the market concentration, results indicate that local market competition is insignificant (p value = 0.127) and negatively associated with audit fees. This is interpreted as finding weak evidence that audit firms situated in competitive local markets offer price discounts in order to entice customers to attain their service. Furthermore, results suggest that local market competition has no impact on audit quality, when using restatements as a

2 Audit firm’s MSA is based on the U.S. Census Bureau’s 2009 classifications (U.S. Department of Commerce, Bureau of Census 2009) published at: http://www.census.gov/population/metro/data/pastmetro.html.

3 MSA and county will be used interchangeable throughout the course of the text.

(7)

proxy for audit quality. Therefore, results indicate that audit firms situated in competitive environments do not reduce audit quality, in response to lower audit fees.

This paper, also evaluates the differentiated effect between Big-4 and Non-Big-4 (hereby referred to as NB4) auditors as although there is domination in national market concentration for Big-4 auditors, the local market shares of individual Big 4 firms varies substantially. It is important to separate Big-4 from NB4 accounting firms, as it is argued that minimal competition exists between the segments (U.S. Senate, 1976; GAO, 2008; TACAP, 2008). This is evident in the pricing premium derived from Big-4 auditors due to increased service and reputation that comes with the audit. Therefore, my hypothesis is that local market competition will have a more pronounced impact on NB4 auditors than Big-4 auditors, after controlling for covariates. Tests indicate this not to be true, indicating that Big-4 auditors compete against NB4 auditors for the service of clients.

Overall, my study empirically addresses four related questions:

i. Does competition exist in highly concentrated markets.

ii. Does the level of local market competition affect the audit fee charged.

iii. Does the level of local market competition impact the audit quality.

iv. Does competition exist between Big-4 auditors and NB4 auditors.

These questions will enable me to comment on the debate surrounding the presence of local audit market competition in highly concentrated audit markets.

A challenge that I will face is when looking at the full sample, results tend to be biased toward audits of higher complexity. To reduce the impact of the bias, I focus part my analysis on Big- 4 and NB4 auditors, as NB4 auditors generally audit clients with lower complexity.

Additionally, I remove counties where annual audit observations are less than 15, to remove counties with weak business environments. Furthermore, separating MSAs (from each state/country) and viewing the competition within each operating county, I am able to measure the local competition presence in each market.

I conduct an additional analysis to further address this challenge, and interpret how prevalent these findings are across my sample. I separate the size of clients into small-medium business’

(hereby referred to as SMB), small-medium enterprises (hereby referred to as SME), and large clients. To which I find evidence that SME clients situated in competitive counties are offered price discounts. Test scores, for SBE clients indicate a positive relationship between audit fee

(8)

and local market competition and large clients indicate a less pronounced negative effect on audit fee compared to that of SME clients. Additionally, I regress audit quality and find similar measures to my original regression.

This study aims to make several important contributions. Firstly, a relevant research topic is addressed that few others have attempted to investigate. Specifically, providing evidence on the impact of local market competition on audit fees and audit quality, of which is important to both regulators and audit market participants, as per GAO (2003, 2008). Secondly, results of this paper will add to the debate on how high market concentration, as alleged by GAO, and competitive markets, can exist simultaneously (Keune, Mayhew & Schmidt, 2016). Finally, by showing the varying level of competition within counties, confirms the importance of measuring at a local level rather than a national level (Kallapur, Sankaraguruswamy, & Zang 2010; Boone, Khurana, & Raman, 2012; Ye, & Zhang, 2018; Eshleman, & Lawson, 2017;

Numan et al., 2012)

The rest of this paper is organised as follows: In the following section, the literature is developed, sections 3 and 4 will discuss the models used, sample selection and empirical test results, as well as the additional tests. Section 5 will provide the concluding remarks.

(9)

2 Literature Review and Hypothesis Development

2.1 Local Market Competition & Audit Fees

Prior to 1987, the top accounting firms were referred to as the Big-8. Later, in 1989, the Big-8 became the Big-6, dropping to the Big-5 in 1998 and further to the Big-4 in 2002 (The Big-4 Accounting Firms, 2021). The evolution of the major accounting firms has triggered regulatory concerns over the current market concentration, which regulators argue has resulted in impediments on the audit market competition (GAO, 2008). It is assumed that the increases in market concentration depict an audit environment associated with lower competition.

However, market concentration is perceived as a static measure of market competition according to industrial organisation theory (Carlton et al., 1994). Despite this, I can infer similarities between market concentration and competition, as a higher market concentration implies fewer suppliers and limited options for clients, which can result in collusion, leading to increased audit fees. Looking at the relationship between market concentration and audit fee’s, there is mixed evidence. Multiple studies report a positive association between local market concentration and audit fee across various kinds of audits; in Canadian municipal audits by non-Big 6 firms (Bandyopadhyay, & Kao, 2004), in Chinese public client audits (Huang, Chang, & Chiou, 2016), and in US public client audits (Eshleman et al., 2017). Furthermore, Gerakos & Syverson (2015) suggested that the loss of a Big-4 firm, which would increase concentration within the audit market, would result in a rise in overall audit fees by an additional $480 – 570 million per year. Therefore I can theorize, that audit markets that suffer from a lack of suppliers (high market concentration), could impede fee competition, following with regulators concerns. Conversely, Pearson & Trompeter (1994) and Numan & Willekens (2012) find that firms located in a more concentrated MSA, pay lower audit fees. Thus, rejecting the concerns expressed by regulators. Although the relationship between market concentration and audit fee’s is beneficial for the understanding of this paper, the literature mentioned above does not specifically use a dynamic measure of competition, in which this paper attempts to explore.

The effect of audit competition on audit fees has been indirectly measured Maher, Tiessen, Colson, & Broman (1992) ,which examines the changes in audit fees over the period 1977 to 1981. This period is considered a time of increasing competition amongst the U.S. audit market and results indicate that audit fees declined notably over this period. In a similar fashion, from

(10)

1985 to 1989, Sanders, Allen, & Korte (1995) found that the municipal audit fees charged, decreased.

Additionally, Ghosh & Lustgarten (2006) find that among the NB4 auditors, initial audit fee discounts are more pronounced among NB4 auditors, indicating that competition exists among the NB4 auditors. These papers deduce that competition leads to lower audit fees, however, they do not provide a direct measure competition and its effect on audit fee. More recently, Van Raak et al. (2020) used a direct measure of audit competition (client mobility) and found that in the Belgian private market, market competition stimulates price competition in both the SME client segment and large-client segment. This paper uses a similar measure to that of Van Raak et al. (2020).

The concern of regulators, about the lack of competition, stems from the acceleration of client growth over recent decades. Client growth gave rise to increased demand for audit services; as a result of increases in size, complexity and a desire for quality by clientele. This follows with the argument that the degree of concentration in the audit market has arisen as a natural response to clients demand, which allowed firms to achieve economies of scale and domination in the market supply (e.g., Danos et al., 1982; Dopuch et al., 1980; Pound et al., 1981; Sirois et al., 2011). Economies of scale provides; advanced audit technologies, lower costs, increased labour, increased clientele, a larger litigation pool and the increased resources required to undertake complex audits (Scherer & Ross 1990; Junius 1997). Therefore, entities are able to charge lower fees to their clients whilst supplying higher audit quality, which enables the auditor to compete for clients and increase their market share. However, if a firm is in an oligopolistic or monopolistic setting, as such the Big-4, fee premiums are likely to be charged, thus uncompetitive pricing behaviour (Yardley, Kauffman, Cairney, & Albrecht 1992). This is supported by economic theory that suggests a positive association between concentration within a local market and fees charged, as higher concentration indicates less suppliers and therefore less competition (Weiss, 1989).

In theory, clients choose their supplier based on their demands for services. The demand is dependent on the specialisation in audit technology (Dopuch et al., 1980; Danos et al., 1982;

Eichenseher., 1984) and reputation that comes with enhanced audit quality (DeAngelo, 1981;

Chow & Rice, 1982; Schwartz & Menon, 1985). The demands differ in multiple features however audit engagement decisions share a common view; clients demand audit services from

(11)

the least cost supplier and only change auditors in response to changes in the amount or type of services required. Put simply, audit firms that have the technology and reputation available to undertake an audit, will compete on price to attain that client.

Therefore, it is reasonable to assume that audit fees are the main source of competition. Based on the above analysis, I provide the following hypothesis;

Hypothesis 1: Ceteris paribus, audit fees will be lower in local markets with higher competition.

2.2 Audit Competition and Audit Quality

Audit committees and Investors desire a higher degree of audit quality since reliable financial statements reduce risk of shareholder litigation, mitigate information risk, increase reputation and lower credit costs (Akerlof, 1970; Jensen & Meckling, 1976; Skinner & Srinivasa, 2012;

Weber et al., 2008; Swanquist & Whited, 2015; Defond & Zhang, 2014; Heininger 2001).

Achieving high quality audits requires the auditor to exert sufficient effort to detect misstatements in the financial reports (DeAngelo, 1981). Auditors must balance the tradeoff between effectiveness and efficiency when choosing the level of audit effort to ensure minimal audit risk and minimal costs. The direct relation between audit fee and audit effort implies that when auditors value the efficiency over the effectiveness of an audit, there is an increased likelihood that audit effort will not be adequate to ensure sound audit quality. Furthermore, in a highly competitive market, auditors must sacrifice audit profit to ensure low audit fees for clients in order to retain their service. Therefore, audit firms are more likely to work towards client retention in competitive markets. Client relations are generally improved by lowering fees, enhancing efficiencies and/or tolerating accounting estimates to their clients (Newton &

Wang, 2013). Without improvements in efficiencies, auditors may allow leniencies in accounting estimates, because providing an alternative opinion may lead to an angered client.

Ultimately, the effect of auditor leniency impairs the independence of the audit, and results in an increased likelihood that errors are not corrected for. Based on the above arguments, I can assume that audit quality will be lower when competition is higher.

On the other hand, Copley & Doucet (1993) found that the number of soliciting bids (competition) for a US governmental audit engagement is positively associated with the

(12)

ultimate quality of the audit, depicting that audit quality actually improves as competition increases. Additionally, Jeter & Shaw (1995) study a market where a ban on direct competition was lifted and found no evidence that competition results in lower audit quality. These papers follow with standard economic theory which argues that competition is positively related to audit quality (e.g., Leland, 1977; Mussa & Rosen, 1978; Spence, 1975). Thus, audit firms provide higher-quality audits, with the aim to build and maintain their reputation with their clients, when competition is higher. For example, Johnstone, Bedard, & Ettredge (2003) found that examining internal data of one audit firm, showed that in a competitive bidding environment the audit firm plans more audit hours while charging lower fees, adhering to the notion that audit quality increases with increased competition. Furthermore, increased competition may result in innovational practices, which could lead to greater efficiencies, improving the audit quality. For example, Deloitte have recently introduced the Automation analysis of SEC and IFRS filing disclosures (Deloitte, 2016). This practice offers real-time, online access to SEC and IFRS filing disclosures and accounting policies, enabling instantaneous peer comparisons as well as updates on emerging disclosures and industry trends.

This allows for increased efficiency and betterments to the audit quality.

Local markets deemed with low competition may also provide enhanced quality audits, since the fee they are able to charge is not constricted by price competition. This enables the auditors to permit more effective audits (i.e., increased time allocated to the audit) and also take a harder line with their clients, as substitutes (auditors) available are lower. Hackenbrack, Jensen, &

Payne (2000) conducted an empirical study in a market where the auditor bidding process excludes fees. The authors document indirect evidence that restrictions on price competition are associated with higher audit quality. However, on the other hand, lower audit quality may exist in local markets associated with low competition, as auditors may become stagnate in their procedures and have minimal desire for improvement⁴. Audit firms may also exercise their market power in a concentrated industry by economizing on audit effort, as has been argued by regulators worldwide. It can also be argued that globalization and the presence of the Big-4 in multiple local markets has minimized these effect, however, branches in less competitive counties may not be allocated adequate resources compared to those in highly competitive local market (e.g., Krishnan, 2005; Choi et al. 2010; Francis et al. 2013).

4 US GAO warns that dominant firms may coordinate actions to convince standard setters to introduce new auditing standards with the sole purpose of generating higher fee income (GAO, 2008)

(13)

It is also possible that the code of ethics, which requires auditors to faithfully adhere to accounting rules and regulations (integrity), keep an objective mind, have confidentiality, and competency, is sufficient enough to ensure that competition does not significantly affect the audit quality in either direction (4 Code of Ethics of Internal Auditors, 2021).

Based on the above analysis, I am able to predict that higher competition will result in contrary effects. I therefore test the following hypothesis:

Hypothesis 2: Ceteris paribus, audit quality will not be influenced by local competition.

2.3 Big-4 Versus Non-Big-4

Prior research has inferred that due to the Big-4 public market concentration and market fee premiums (Hay, Knechel, & Wong 2006), little competition exists between the Big-4 and NB4 auditors (U.S. Senate 1976; GAO 2008; TACAP 2008). When a client switches from an incumbent auditor to a competitor, the competitor is required to hold the capacity, such as technology, expertise, and staff, to engage the oncoming client. NB4 auditors lack the economies of scale, in comparison to Big-4 auditors, required to easily accept new clients, as the local audit office’s current staff will be employed to service the engagements at hand, due to the profit maximising theory. In turn, Big-4 auditor offices, will have greater elasticity in resources to deal with potential new demand, evidenced by their larger employee pool, increased size and technologies available. As a result, Big-4 auditors, that are classified as having larger operations, are likely to be better equipped than smaller NB4 competitors, with respect to the ability to facilitate new clients. This results in Big-4 audit firms having an advantage in their ability to facilitate new clients in comparison to NB4 firms, whom will generally need to sacrifice profit margins, in order to compete for the service. Furthermore, it is well documented that NB4 firms have inferior audit quality (Hrazdil, Simunic, &

Suwanyangyuan, 2020) and therefore generally need to compromise on audit fees in order to entice the client in to choosing their services.

Adding to this notion, clients of small audit firms can switch to alternative auditing firms operating in the market at a relatively low transaction cost, due to the client specific knowledge.

Conversely, large firms in the audit market incur higher transactions costs in attempting to

(14)

switch to alternative smaller suppliers (Chu, Simunic, Ye, & Zhang, 2018). Therefore, Big-4 firms in the local market have a competitive advantage in client retention and improving their market concentration, as their elasticity to increase clientele base exceeds NB4 firms. This can be seen evidently in practice, as although all auditors are permitted to audit all firms, the Big- 4 firms have considerable market supply domination.

Big-4 auditors also are accredited with superior attributes, such as being more specialised and reputable (Cranswell, Francis, & Taylor, 1995; Dutillieux, Stokes, & Willekens, 2013). These attributes allow Big-4 firms to charge fee premiums for compensation (Hay, et al., 2006).

Keune et al. (2016), suggests that the NB4 leaders have the potential to be substitutes for the Big-4 firms, however the standard Big-4 premium exceeds the NB4 market leader premium, resulting in a less pronounced effect on the pricing power in the presence of competition compared for NB4 auditors. In following, the GAO (2008), documented that large companies choose Big-4 audit firms as their auditors due to their general capabilities and industry specialized expertise (Hrazdil, et al., 2020).

Lastly, Defond & Zhang (2014) contend that larger audit firms have greater incentives (e.g.

reputation loss from low audit quality audit) to supply high quality audits and are better equipped (e.g. enhanced audit inputs, industry expertise) to deliver high quality audits.

Therefore, I can predict the following two hypotheses:

Hypothesis 3: Ceteris paribus, local markets with higher competition; will have a more pronounced effect on the audit fee charged for NB4 auditors than Big-4 auditors.

Hypothesis 4: Ceteris paribus, local markets with higher competition; will have a more pronounced effect on the audit quality of NB4 auditors than Big-4 auditors.

(15)

3 Sample and Methodology

3.1 Sample Selection

I obtain data from Compustat and Audit Analytics. Table 1, presents the sample selection procedures. The initial sample, which I label Sample A for ease of reference in Table 1, consists of 104,680 client-year observations of US audit engagements for the fiscal years 2011 to 2019.

As summarized in Table 1, I delete 45,039 observations from the sample with lagged or missing total assets (measuring market concentration), 2,622 observations for which the log of assets is less than 0, 13,753 observations for which client location data is missing, and 844 observations for which the client is located outside the US. Exclusion of these observations results in a sample of 62,568 observations, which I will refer to as Sample B. I use Sample B to compute the audit market competition and structure measures.

When calculating measures of control for audit fee and audit required required for my regressions, I exclude 11,581 observations pertaining to counties that have less than 15 observations per annual year, 3,049 observations from 2010, because competition requires a lagged year effect, 6,656 observations of financial and utility institutions, because of their specific audit requirements and accounting procedures, as well as 2,553 observations with extreme changes in total assets⁵. Finally, I remove 2,273 observations without the necessary data to estimate the regression (Equation 3). The resulting pooled sample contains 15,513 observations spanning 2011 – 2019, which I label as Sample C.

TABLE 1 Sample Selection

No. of Obs.

A) Initial Sample Less:

104,680 Observations with missing total assets (45,039)

Observations if log of assets<0 (2,662)

Observations with missing coordinates (13,753)

Observations if currency is not in USD (844) (62,568)

B) Sample used to compute measures of Market Structure 42,112 Less:

Observations where County < 15 Observations (11,581)

Observations in 2010 (3,049)

Utility and Financial Institutions (6,656)

Observations with extreme changes in total assets³ (2,553)

Observations with missing financial information (2,273) (26,599)

C) Sample used to compute Regression 15,513

5 Companies with extreme changes in firm size are classified by companies for which total assets increased by more than 100% or decreased by more than 50%

(16)

3.2 Market Competition Measure

Based off industrial organisation theory, Carlton and Perloff (1994) argue that seller concentration is a static measure of market competition which may represent a part of competition but not the rivalry that exists among audit suppliers in the market. Thus, to measure local market competition, a similar measure following Caves and Porter (1978), Buijink et al.

(1998) and Van Raak et al. (2020), is employed. My paper utilities the market client mobility measure to obtain an estimation for the local market competition. I employ the market share measure (1) to obtain the market competition measure (2).

𝑀𝑎𝑟𝑘𝑒𝑡 𝑆ℎ𝑎𝑟𝑒 = ∑_𝑖=1[𝐼𝑛(𝐴𝑠𝑠𝑒𝑡𝑠_𝑖) × 𝐷_𝑖

∑_𝑖=1[𝐼𝑛(𝐴𝑠𝑠𝑒𝑡𝑠_𝑖)

where ln(Assetsi) is the natural log of total assets of client 𝑖 in market k, Di is an indicator variable that is equal to one if audit firm l audits client 𝑖’s financial statements, and 𝑖 is the total number of clients in market k and year t.

Conceptually, audit fees are the favorable choice to calculate market share (1). Despite this, this paper utilizes total assets to measure market share, due to the assumption that firm size, as measured by total assets, is the primary driver of audit fees. The log-transformation of total assets is used to account for the non-linearity in the total assets–audit fee relationship.

Following Caves and Porter (1978) and Buijink et al. (1998), Raak et al. (2020), I use the instability of audit firms’ market shares as a positive measure of competition. This measure, labelled 𝐿𝑜𝑐𝑎𝑙 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑜𝑚𝑝𝑒𝑡𝑖𝑡𝑖𝑜𝑛_𝑘𝑡 is calculated as the sum of the absolute values of the annual percentage-point changes in market share for each audit firm in a local audit market (2).

𝐿𝑜𝑐𝑎𝑙 𝑀𝑎𝑟𝑘𝑒𝑡 𝐶𝑜𝑚𝑝𝑒𝑡𝑖𝑡𝑖𝑜𝑛_𝑘𝑡 = ∑ ∣ 𝑀𝑎𝑟𝑘𝑒𝑡 𝑆ℎ𝑎𝑟𝑒_{𝑙𝑘,𝑡}− 𝑀𝑎𝑟𝑘𝑒𝑡 𝑆ℎ𝑎𝑟𝑒_{𝑙𝑘,𝑡−1} ∣ (1)

(2)

(17)

3.3 Restatements as a Proxy for Audit Quality

Audit Quality has multiple variations of measurement. I focus on restatements as a proxy for audit quality as restatements measure a failure in the primary responsibility of the auditor. An auditor’s responsibility requires them to deliver financial statements free of material misstatement or error. Therefore, restatements indicate that auditors have failed to act dutifully in their responsibilities to supply financial statements free of error (Newton et al., 2013, Palmrose & Scholz, 2004). This can reduce the public trust investors hold towards entities, signifying poor quality of financial reporting. Past literature has regularly employed restatements to measure the audit quality possessed by clients (e.g., Kinney et al., 2004;

Schmidt, 2012; Schmidt and Wilkins, 2013; Newton et al., 2013). It is inferred that there is a required audit effort necessary to determine the material accuracy of the financial statements of a client. Amongst clients, the amount of audit effort required to undertake an accurate audit differs, due to the financial statement preparation of the client. In practice, it is an auditors duty to ensure that the audit effort is sufficient to ensure a high quality audit to support the opinion expressed by the auditor (Palmrose and Scholz., 2004). Generally, if an unqualified opinion on misstated financial statements occurs, it is likely that auditor’s effort was inadequate to derive the audit opinion (Francis & Michas, 2013).

It must be noted that no specific measure of audit quality is not without measurement error.

Nonetheless, using restatements as a measure of audit quality allows appropriate conclusions to be drawn, however, results should be approached with caveat.

3.4 Regression Models

To examine the effect of competition on audit fee and audit quality and test my hypothesis, I will estimate the following regression equation (3):

𝐿𝑁𝐴𝐹_𝑖𝑡 𝑜𝑟 |𝑅𝐸𝑆𝑇𝐴𝑇𝐸𝑀𝐸𝑁𝑇𝑆|_𝑖𝑡 = 𝛽₀ + 𝛽₁𝐴𝑈𝐷𝐼𝑇_𝐶𝑂𝑀𝑃 + 𝛽₁𝐵𝐼𝐺4 +

∑𝛽_𝑧𝐶𝑂𝑁𝑇𝑅𝑂𝐿𝑆 + ∑ 𝑌𝑒𝑎𝑟 + ∑ 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 + ∑ 𝑀𝑆𝐴_𝑘+ ℇ

The dependent variable (𝐿𝑁𝐴𝐹_𝑖𝑡) is the natural logarithm of client i’s audit fees (in $ thousands) in year (t), and (|𝑅𝐸𝑆𝑇𝐴𝑇𝐸𝑀𝐸𝑁𝑇𝑆|_𝑖𝑡) identifies firms that have financial restatements. My

(3)

(18)

test variable of interest is (AUDIT_COMP) which is the level of auditor competition within a local market, which is described in section 3.2. (CONTROLS) is the a vector of variables that control for client and auditor characteristics, which is further discussed for LNAF in section 3.5.1 and 𝑅𝐸𝑆𝑇𝐴𝑇𝐸𝑀𝐸𝑁𝑇𝑆 in section 3.5.2. The model also clusters standard errors by firms⁶ because the sample includes multiple observations per client, which could cause potential cross sectional dependence. Additionally, I control for (∑𝑌𝑒𝑎𝑟 ) which controls for that year, the industry fixed effects (∑𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦) as described by Fama and French (1997), and the MSA fixed effects (∑ 𝑀𝑆𝐴_𝑘) as described by Eshleman & Lawson (2017). I further include the dummy variable (BIG4) in order to test hypothesis 3 and 4, which focuses on the impact of Big-4 versus NB4 auditors. I define and discuss my control variables below.

3.5 Control Variables

3.5.1 Audit Fee Control Variables

𝑋^′𝛽 = 𝛽₂𝑆𝐼𝑍𝐸 + 𝛽₃𝐶𝐴 + 𝛽₄𝐼𝑁𝑉𝑅𝐸𝐶 + 𝛽₅LEV + 𝛽₆𝑅𝑂𝐴 + 𝛽₇𝐿𝑂𝑆𝑆 + 𝛽₈𝐶𝑅 + 𝛽₉𝐵𝑇𝑀 + 𝛽₁₀𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽₁₁𝐵𝑈𝑆𝑆𝐸𝐺 + 𝛽₁₂𝐺𝐸𝑂𝑆𝐸𝐺 + 𝛽₁₃𝐼𝑆𝑆𝑈𝐸 + 𝛽₁₄𝐸𝑋𝑂𝑅𝐷 + 𝛽₁₅𝐹𝑂𝑅𝐸𝐼𝐺𝑁 + 𝛽₁₆𝑃𝐸𝑁𝑆𝐼𝑂𝑁𝑆 + 𝛽₁₇𝑁𝐴𝑇_𝑆𝑃𝐸𝐶 + 𝛽₁₈𝐶𝐼𝑇𝑌_𝑆𝑃𝐸𝐶 + 𝛽₁₉𝐽𝑂𝐼𝑁𝑇_𝑆𝑃𝐸𝐶 + 𝛽₁₆𝐼𝐶_𝑊𝐸𝐴𝐾 + 𝛽₁₆𝐿𝐼𝑇

I am required to control for auditor characteristic variables that impact the audit fee. Hay et al., 2006; Desir et al. 2014; Eshleman and Guo 2014a, attribute four factors that affect the audit fee charged. These factors are associated with measures of client size, client risk, client complexity, and auditor-and-engagement-specific attributes.

To capture the size of the client (SIZE), the natural log of total assets is calculated, as it can be expected that larger clients are charged higher fees. For the second factor, client risk, I include multiple measures, explained below, as I expect a positive relationship between audit fee and risk of a client (Simunic & Stein, 1996). The client’s proportion of current assets (CA), the

6 Results are robust to clustering standard errors by audit firm.

(4)

(19)

current ratio (CR), the performance (ROA), are included to capture any effects that a clients profitability has on the fee charged, as less financially stable clients are likely to be charged heightened fees. Additionally, I include (LOSS) as an indicator variable equal to 1 for firms with negative income before extraordinary items in the current year, inventory/receivables (INVREC) and leverage (LEV) which is defined as long term debt divided by total assets. I also include, (LIT) as an indicator variable set to 1 if the firm operates in an industry associated with high litigation because prior studies indicate that litigation environments are linked to higher audit fees (Fuerman 1997; Raghunandan et al. 2003). Furthermore, I also include the presence of internal control weaknesses (IC_WEAK), because firms with internal control weaknesses generally have poor accounting systems in place and thus requires heightened effort from auditors Finally, I include two proxies for client growth (GROWTH and BTM) because growth firms are charged higher audit fees (Choi, et al., 2010; Lobo & Zhao 2013).

When measuring client risk I expect the coefficients to reflect higher audit fees charged for riskier clients.

To capture client complexity, the following variables are included; a client’s number of business segments (BUSSEG) and geographic segments (GEOSEG), whether the client issued securities during the year (ISSUE), whether the client had extraordinary items (EXORD) or foreign income (FOREIGN), and whether the client had a pension fund during the year (PENSION). I expect the coefficients of all variables to be positive, inferring that the more complex clients are the higher the audit fees (Casterella, Francis, Leis, and Walker, 2004; Hay, Knechel, and Wong 2006).

The fourth factor that affects audit fees is auditor and engagement attributes. To control for premiums for industry specialization I include both measures of CITY_LEADER, NAT_LEADER, JOINT_SPEC; (Francis et al., 2015).

All continuous firm-level control variables, with the exception of BUSSEG and GEOSEG, are winsorized at the first and 99th percentiles.

(20)

3.5.2 Audit Quality Control Variables

𝑋^′𝛽 = 𝛽₂𝑂𝐹𝐹𝐼𝐶𝐸_𝑆𝐼𝑍𝐸 + 𝛽₃𝐶𝐼𝑇𝑌_𝐿𝐸𝐴𝐷𝐸𝑅 + 𝛽₄𝑁𝐴𝑇_𝐿𝐸𝐴𝐷𝐸𝑅 + 𝛽₅AUDIT_FEES + 𝛽₆𝐹𝐸𝐸_𝑅𝐴𝑇𝐼𝑂 + 𝛽₇𝐼𝑁𝐹𝐿𝑈𝐸𝑁𝐶𝐸 + 𝛽₈𝑄𝑈𝐴𝐿_𝑆𝐼𝑍𝐸 + 𝛽₉𝐿𝐸𝑉 + 𝛽₁₀𝐴𝐶𝐶𝐸𝐿_𝐹𝐼𝐿𝐸𝑅 + 𝛽₁₁𝐶𝐴

+ 𝛽₁₂𝐿𝑂𝑆𝑆 + 𝛽₁₃𝑅𝑂𝐴 + 𝛽₁₄𝐿𝐼𝑇 + 𝛽₁₅𝐼𝑁𝑉𝑅𝐸𝐶 + 𝛽₁₆𝐺𝑅𝑂𝑊𝑇𝐻 + 𝛽₁₇𝑀𝐸𝑅𝐺𝐸𝑅 + 𝛽₁₈𝑅𝐸𝑆𝑇𝑅𝑈𝐶𝑇 + 𝛽₁₉𝐼𝐶_𝑊𝐸𝐴𝐾

I am required to control for auditor and client characteristics when measuring audit quality.

Recent studies have found that larger offices conduct higher-quality audits (e.g., Francis et al., 2009; Choi et al., 2010; Francis et al., 2013), therefore OFFICE_SIZE is controlled for, which is calculated as the log of total audit assets charged by the office during year t.

Similarly, as audit fee control variables, CITY_LEADER and NAT_LEADER controls are included prior literature finds that industry expertise improves audit quality (e.g., Krishnan, 2003; Francis et al., 2005; Reichelt, & Wang, 2010). I also control for AUDIT_FEES, which is the log of the audit fees paid by the client firm in year t, FEE_RATIO, which is the firm’s payments for non-audit fees divided by total audit fees paid (audit plus non-audit fees), and leverage (LEV) because the economic bonding that occurs between the client and the auditor creates two effects. (1) a large client may have more leverage with its auditor (Francis et al., 2013), and (2) restatements are positively associated with unspecified non-audit services (Kinney et al., 2004). Control variables are also included to capture risk characteristics of the client that might affect the likelihood of restatement. Larger clients generally develop enhanced control systems compared to smaller clients, therefore I expect that larger clients will be less likely to misstate financial statements (QUAL_SIZE). Furthermore, I include the client’s proportion of current assets (CA), the performance (ROA), to capture any effects that a clients profitability has on the level of restatement, as less financially stable clients are more likely to overstate earnings. Additionally, I include loss (LOSS) as an indicator variable equal to 1 for firms with negative income before extraordinary items in the current year inventory, and receivables (INVREC), which is defined as long term debt divided by total assets. Furthermore I include, (GROWTH), because growth firms ordinarily experience increased restatements, (ACCEL_FILER) as an indicator variable equal to 1 for firms that are accelerated filers, and (LIT) as an indicator variable set to 1 if the firm operates in a litigation-risky industry, as it has

(5)

(21)

been documented that firms situated in risky environments experience higher restatements (Raghunandan et al., 2003). Finally, I control for (MERGER) and (RESTRUCT), because firms undergoing growth generally suffer from accounting changes, resulting in internal control weaknesses (e.g., Ashbaugh-Skaife, LaFond, & Mayhew, 2007; Richardson, Tuna, & Wu, 2002). I additionally include (IC_WEAK) to control for the presence of internal control weaknesses. My expectation is that this coefficient will be positive. All continuous firm-level control variables are winsorized at the first and 99th percentiles.

Attached in Appendix A is a detailed definition for each of the variables I have chosen to include.

(22)

4 Results

4.1 Descriptive Statistics Audit Fee

Table 2, displays 15,513 observations of descriptive statistics for audit fees (AFEE), local market competition (MRKT_COMP), local competition (LOC_COMP), Herfindahl Index at local market level (MRKT_CONCEN), Herfindahl Index at local level (LOC_CONCEN), control variables and the differentiated effect between Big-4 firms (BIG4) and NB4 firms (NB4), displayed in columns 7 and 8. All continuous independent variables, with the exception of BUSSEG and GEOSEG, are winsorized at the 1st and 99th percentile to reduce the influence of outliers. The mean and median values of the log transformation of audit fees, is 13.895 and 13.998, respectively, as is consistent with previous literature. While the average log of audit fees amount to 13.895, there is substantial variation in audit fees, which most likely reflects the sample variation in audit complexity, which will be discussed in the additional tests, in section 4.5. Local market competition and market competition values are on average, 0.075, and 0.181, respectively. These values are lower compared to local competition (client mobility) levels observed in prior research focusing on the private markets of Belgium (Raak et al. 2020), Germany and the Netherlands (Buijink et al., 1998). The local market level of competition (MRKT_COMP) indicates that auditors lose an average, annual total of 3.75% (0.075/2) of local market share to their competitors. Furthermore, the local competition level (LOC_COMP) implies that clients lose on average 9% (0.183/2) of total local clients each year. Looking at both local competition and local market competition, the variation in variables is substantially high. This infers that competition varies significantly. This can be expected due to the geographical differences explored throughout U.S. counties.

CONCEN, which is based on the local-level Herfindahl index, has a mean (median) value of 0.196 (0.184), consistent with research in this area (e.g., Boone et al., 2012). Additionally, IND_CONCEN, which is based on the local market-level Herfindahl Index, has a mean (median) value of 0.092 (0.00). By separately identifying the measures, the importance of looking at local market level competition rather than primarily at the local level competition is evident. The importance stems from the right skewed metric. The IND_CONCEN median value of 0, and mean of 0.092, indicates either; a market where the volume of audits in a particular industry is low and hence one auditor holds majority of the market share, or that majority of audit firms do not compete for services they are unfamiliar with. Thus, firms with local market hold, will generally influence majority of the local market share, which could

(23)

result in impediments on audit fee pricing competition. Adhering to the notion, results indicate lower competition experienced within local markets (0.075) in comparison to the total local competition (0.183).

The last two columns of table 2 show the difference in mean values between Big4 and NB4 auditors. It is observed that on average, Big-4 auditors receive higher audit fees, experience lower competition, are larger in size (complexity) and are more likely to be city/national specialist auditors.

TABLE 2: Descriptive Characteristics for Audit Fee measure (n = 15,511) BIG4

( n = 10,716)

NB4

(n = 4,795)

Variables Mean SD P25 Median P75 Mean Mean

AFEE 13.895 1.328 13.060 13.998 14.779 14.483 12.580

MRKT_COMP ^0.075 ^0.109 ^0.005 ^0.031 ^0.093 ^0.071 ^0.083

LOC_COMP 0.183 0.134 0.082 0.144 0.241 0.179 0.192

MRKT_CONCEN ^0.068 ^0.184 ^0.000 ^0.000 ^0.038 ^0.071 ^0.063

LOC_CONCEN ^0.196 ^0.053 ^0.163 ^0.184 ^0.217 ^0.202 ^0.181

SIZE ^6.349 ^2.264 ^4.824 ^6.453 ^7.923 ^7.294 ^4.239

CA ^0.492 ^0.253 ^0.294 ^0.482 ^0.689 ^0.463 ^0.558

INVREC ^0.238 ^0.178 ^0.093 ^0.205 ^0.343 ^0.216 ^0.287

LEV 0.263 0.276 0.030 0.212 0.386 0.267 0.255

ROA ^-0.068 ^0.296 ^-0.077 ^0.024 ^0.068 ^-0.017 ^-0.181

LOSS 0.393 0.488 0.000 0.000 1.000 0.320 0.555

CR 2.704 2.554 1.258 1.941 3.123 2.620 2.891

BTM 0.454 0.838 0.177 0.375 0.693 0.436 0.495

GROWTH ^0.138 ^0.519 ^-0.034 ^0.058 ^0.176 ^0.128 ^0.160

BUSSEG ^1.118 ^0.271 ^1.000 ^1.000 ^1.099 ^1.142 ^1.066

GEOSEG ^1.193 ^0.445 ^1.000 ^1.000 ^1.386 ^1.232 ^1.105

ISSUE ^0.816 ^0.388 ^1.000 ^1.000 ^1.000 ^0.849 ^0.740

EXORD 0.148 0.356 0.000 0.000 0.000 0.162 0.118

FOREIGN 0.656 0.475 0.000 1.000 1.000 0.746 0.456

PENSIONS 0.925 0.264 1.000 1.000 1.000 0.948 0.873

NAT_SPEC 0.066 0.248 0.000 0.000 0.000 0.095 0.000

CITY_SPEC 0.193 0.395 0.000 0.000 0.000 0.241 0.086

JOINT_SPEC ^0.029 ^0.169 ^0.000 ^0.000 ^0.000 ^0.043 ^0.000

IC_WEAK ^0.078 ^0.269 ^0.000 ^0.000 ^0.000 ^0.347 ^0.336

LIT ^0.344 ^0.475 ^0.000 ^0.000 ^1.000 ^0.044 ^0.156

Notes: This table presents the descriptive analysis for all Audit Fee variables used in the regression model for the purpose of this study. All variables are defined in Appendix A. The full sample consists of 15,513 firm- year observations during the period 2011-2019 for 74 U.S. Metropolitan Statistical Areas. All continuous variables are winsorized by fiscal year at the 1st and 99th percentile. Total assets and sales are in millions.

(24)

For brevity, the remaining descriptive statistics for the sample variables are not discussed, as they appear consistent with prior studies of audit fees (Francis et al., 2005; Ghosh et al., 2006;

Dunn et al., 2011; Numan et al., 2012; Eshleman et al., 2017).

Table 3, presents descriptive statistics for the ten least competitive and highest competitive local audit markets. Since the local market competition level and concentration measure changes every year, the mean across all years in the sample is utilized to determine the values.

The population for the MSA’s per the 2019 U.S. Census, as well as the 2019 median accountant and auditor wage for each MSA⁷ is included to highlight the importance of controlling for MSA fixed effects. As reported, the most competitive audit markets include MSA’s such as San Mateo, CA, Santa Clara, CA, Boulder, CO, and Harris, TX.

7 The wage data shown here are available at the Bureau of Labor Statistics website:

https://www.bls.gov/oes/current/oes132011.htm

TABLE 3: Descriptive Characteristics for MSA (n = 15,513)

MSA Name

All Audit Firms in MSA

Big-4 Audit Firms in MSA

Population

Median Auditor Salary

Mean Local Comp

Mean Herfinda hl Index

Mean Audit Firms

Mean Local

Comp % of clients Mean Audit Firms

Ten Least Competitive MSAs

Franklin, OH 1,317,000 58,900 0.008 0.319 5.82 0.002 83.9% 4.00 Delaware, PA 566,747 56,500 0.022 0.310 5.68 0.012 85.2% 3.16 Wayne, MI 1,749,000 56,120 0.025 0.274 5.03 0.019 85.7% 3.00 St. Louis, MO 996,919 53,200 0.026 0.243 6.55 0.019 88.3% 4.00

Denton, TX 887,207 61,150 0.027 0.222 7.52 0.009 46.6% 2.73

Cuyahoga, OH 1,235,000 22

53,090 0.028 0.308 7.67 0.010 82.5% 4.00 Bexar, TX 2,004,000 55,610 0.029 0.324 5.25 0.010 63.3% 2.52 Ventura, CA 846,006 58,570 0.029 0.320 6.99 0.013 63.4% 2.84 New York, NY 8,419,000 69,680 0.030 0.160 26.28 0.009 65.7% 4.00 Mecklenburg, NC 1,110,000 59,510 0.030 0.296 5.77 0.010 89.3% 4.00 Ten Most Competitive MSAs

San Mateo, CA 4,731,803 67,190 0.190 0.248 11.49 0.115 78.2% 4.00 Boulder, CO 326,196 58,890 0.137 0.173 6.52 0.125 30.7% 2.63 Santa Clara, CA 1,990,660 68,800 0.136 0.198 15.12 0.033 77.3% 4.00 Harris, TX 4,713,000 48,040 0.112 0.165 18.39 0.046 67.6% 4.00 Travis, TX 1,274,000 56,080 0.107 0.218 13.02 0.038 56.3% 3.59 Norfolk, MA 706,775 62,510 0.106 0.201 6.69 0.066 67.8% 3.89 Palm Beach, FL 1,497,000 52,570 0.099 0.177 13.04 0.053 36.2% 2.41 Monmouth, NJ 618,795 69680 0.092 0.175 7.04 0.012 30.7% 2.00 Wake, NC 1,112,000 56,610 0.092 0.233 6.20 0.027 59.1% 3.32 Middlesex, NJ 825,062 62,510 0.091 0.180 12.78 0.166 51.3% 3.32 Notes: This table presents the descriptive statistics for the ten least and most competitive MSAs in my pooled sample. Data for the number of audit firms comes from the Audit Analytics database. Population and auditor wage data are for the year 2019 and come from the 2019 U.S. Census and the Bureau of Labor Statistics, respectively.

(25)

As tabulated, MSA’s with higher competition, generally, have more audit firms for clients to choose from, have lower MSA Herfindahl Index’s, have more NB4 presence and local market competition is less prevalent for Big-4 auditors than NB4 auditors. This reinforces the importance of separating Big-4 and NB4 auditor segments, indicating Big-4 firms experience less local market competition. Overall, Table 3 indicates preliminary evidence that market competition is more pronounced for NB4 firms.

Table 4 displays Pearson correlations among audit fees, local market competition, local competition, Herfindahl Index at the local market level, Herfindahl at local level, and control variables. The univariate correlation between AFEE and SIZE is positive and economically significant. The correlation between LOC_CONCEN and MRKT_COMP is negative.

Additionally the correlation between LOC_CONCEN and SIZE is positive. However, the correlation between MRKT_COMP and SIZE is negative. Thus, holding with Raak et al., 2020, the MRKT_COMP measures a separate dimension of market structure.

Additionally, I am able to view a positive correlation between AFEE and the LOC_CONCEN, indicating that counties situated in higher concentrated areas, do in fact receive higher audit fees. Furthermore, in viewing the negative correlation between AFEE and MRKT_COMP, I see preliminary evidence that in the presence of local market competition, audit fees are lower.

(26)

(27)

4.2 Descriptive Statistics Audit Quality

Table 2, displays 15,513 observations of descriptive statistics of audit quality, local market competition, local competition, Herfindahl Index at local market level, Herfindahl at local level, control variables and the differentiated effect between Big-4 firms and NB4 firms, displayed in column 7 and 8. All continuous independent variables, are winsorized at the 1st and 99th percentile to reduce the influence of outliers. I conduct my analysis using the initial definition of restatements (i.e., each restatement is allocated to all affected periods).

The mean and median values of restatements is approximately 0.070 and 0, indicating that the likelihood of a restatement occurring is 7%.

For brevity, I do not to discuss the descriptive statistics for the audit quality sample variables as they appear consistent with prior studies of audit quality (e.g., Kinney et al. 2004; Choi et al. 2010; Reichelt & Wang 2010; Dunn et al. 2011; Francis et al. 2013; Newton et al. 2013).

The last two columns of table 5 show the difference in means between Big-4 and NB4 auditors.

I observe that, on average Big-4 auditors receive higher audit fees, experience lower competition, are larger in size (complexity) and are more likely to be city/national specialist auditors. The variable of interest, restatements is higher for Big-4 auditors than is for NB4 auditors. This violates the theory that Big-4 auditors provide enhanced service and provide better quality audits. However, the variation between OFFICE_SIZE for Big-4 and NB4 auditors indicates this is most likely due to the complexity of the audit, (Big-4 = 16.973; NB4 auditor, 13.341) which will be discussed in section 4.5.

Table 6, presents a Pearson correlations table to view the associations between variable pairs.

I do not discuss Table 6 in depth, as most significant findings were discussed in Table 4. The preliminary findings in Table 6 indicate that local market competition has a small negative correlation with audit quality. I examine these relationships further in section 4.4.

(28)

BIG4

( n = 10,716)

NB4

(n = 4,795)

Variable Mean SD P25 Median P75 Mean Mean

RESTATEMENT ^0.070 ^0.255 ^0.000 ^0.000 ^0.000 ^0.081 0.046

MRKT_COMP ^0.076 ^0.113 ^0.005 ^0.031 ^0.093 ^0.071 0.085

LOC_COMP ^0.183 ^0.134 ^0.082 ^0.144 ^0.241 ^0.179 0.192

MRKT_CONCEN 0.068 0.184 0.000 0.000 0.038 0.071 0.064 LOC_CONCEN ^0.196 ^0.053 ^0.163 ^0.184 ^0.217 ^0.202 0.181 OFFICE SIZE ^15.883 ^2.065 ^14.568 ^16.319 ^17.474 ^16.973 13.431 CITY LEADER 0.193 0.395 0.000 0.000 0.000 0.241 0.086 NAT LEADER 0.066 0.248 0.000 0.000 0.000 0.095 0.000 AFEE ^13.895 ^1.328 ^13.060 ^13.998 ^14.779 ^14.486 12.567

FEE RATIO 0.097 0.125 0.000 0.042 0.163 0.127 0.035

INFLUENCE 0.303 0.339 0.049 0.148 0.436 0.177 0.585

Q_SIZE 6.111 2.342 4.547 6.296 7.734 7.010 4.099

LEV ^0.263 ^0.276 ^0.030 ^0.212 ^0.386 ^0.268 0.332

ACCEL FILER 0.046 0.210 0.000 0.000 0.000 0.054 0.028

CA 0.492 0.253 0.294 0.482 0.689 0.463 0.558

LOSS 0.393 0.488 0.000 0.000 1.000 0.320 0.555

ROA -0.068 0.296 -0.077 0.024 0.068 -0.016 -0.239

LIT 0.344 0.475 0.000 0.000 1.000 0.347 0.338

INVREC 0.238 0.178 0.093 0.205 0.343 0.216 0.289

GROWTH 0.138 0.519 -0.034 0.058 0.176 0.274 0.595

MERGER 0.409 0.492 0.000 0.000 1.000 0.486 0.235

RESTRUCT 0.642 0.479 0.000 1.000 1.000 0.569 0.806

IC WEAK 0.078 0.269 0.000 0.000 0.000 0.044 0.158

Notes: This table presents the descriptive analysis for all Audit Quality variables used in the regression model for the purp ose of this study. All variables are defined in Appendix A. The full sample consists of 15,513 firm-year observations during the period 2011-2019 for 47 U.S. Metropolitan Statistical Areas. All continuous variables are winsorized by fiscal year at the 1st and 99th percentile.

(29)

The Impact of Local Competition on Audit Fees and Audit Quality.

The Impact of Local Competition on Audit Fees and Audit Quality.

Contents