Voluntary Audits Versus Mandatory Audits 3qj5h

See discussions, stats, and author profiles for this publication at: https://idoc-pub.futbolgratis.org/publication/228273267

Voluntary Audits Versus Mandatory Audits Article  in  The ing Review · August 2011 DOI: 10.2308/accr-10098

CITATIONS

READS

66

971

2 authors: Clive S. Lennox

Jeffrey Pittman

Nanyang Technological University

Memorial University of Newfoundland

56 PUBLICATIONS   2,442 CITATIONS   

48 PUBLICATIONS   1,291 CITATIONS   

SEE PROFILE

All content following this page was ed by Clive S. Lennox on 11 November 2014. The has requested enhancement of the ed file.

SEE PROFILE

Voluntary Audits versus Mandatory Audits*

Clive Lennox** Nanyang Technological University, Singapore 639798 [email protected]

Jeffrey Pittman Memorial University of Newfoundland, St. John’s, NL A1B 3X5, Canada [email protected]

September 2009

*

We appreciate helpful comments from Elisabeth Dedman, Omrane Guedhami, Gilles Hilary, Siew-Hong Teoh, Nan Zhou, and seminar participants at Nanyang Technological University on an earlier version of this paper.

**

Corresponding author: [email protected].

1

Voluntary Audits versus Mandatory Audits

Abstract An important policy question is whether audits of financial statements should be mandatory or voluntary. A major economic rationale for mandatory auditing is that financial statement s enjoy benefits from having access to reliable audited information for every company. On the other hand, advocates for voluntary auditing argue that imposing audits removes the signaling value stemming from companies revealing whether they wish to be audited. To evaluate these arguments empirically, we examine a setting in which mandatory audits were replaced by voluntary audits for private UK companies. We first demonstrate that the benefits from forcing companies to be audited are modest because companies that do not want to be audited choose low quality audit firms and pay lower audit fees during the mandatory regime. Second, we show that there is an important signaling benefit when companies are allowed to choose whether they will be audited. After mandatory audits are replaced by a voluntary regime, companies that retain their auditors enjoy significant upgrades to their credit ratings. In contrast, companies that dispense with being audited suffer downgrades to their credit ratings since avoiding an audit sends a negative signal and their unaudited financial statements are less credible.

2

1. Introduction Requiring independent audits is an important policy mechanism available to governments to regulate the supply of reliable ing information to investors (Barton and Waymire, 2004). Information is a public good and, as with all public goods, there is a concern that too little would be supplied under private contracting.1 According to this market failure argument, companies should be compelled to have their financial statements audited to ensure that outsiders can have access to reliable information. In the other direction, the ers of voluntary audits stress that the mandatory requirement suppresses the signal that is conveyed when companies can choose whether to be audited (Sunder, 2003). The purpose of this study is to provide empirical evidence on the merits of these competing arguments by analyzing outcomes stemming from voluntary versus mandatory audits. External audits are imposed on all publicly traded companies in the US and UK, although these countries differ in their policies towards private companies. In contrast to the US where audits of private companies remain voluntary, they were mandatory in the UK until recently.2 In 1994, the EU Fourth Directive permitted national governments to abandon the requirement that small companies submit to an audit. A subsequent Firms operating in an unregulated environment understandably focus on their own costs and benefits without considering the socially optimal level of disclosure. Zingales (2009: 394) stresses that firms disclose sub-optimally because firm-level benefits from disclosure are smaller than its society-level benefits: “General Motors’ disclosure helps investors evaluate Ford, but GM will never internalize this benefit.” Since the private and social values of information can diverge, regulation is frequently justified on the grounds that this induces positive externalities (e.g., Dye, 1990; ati and Pfleiderer, 2000; Lambert et al., 2007). However, other theory implies that, rather than maximizing social welfare, mandating disclosure of ing information can generate negative externalities (e.g., Fishman and Hagerty, 1989). Leuz and Wysocki (2008) survey this literature. 1

2 We do not examine the effect of the 1994 change in regulation because our data source provides financial statement information for the most recent ten years only. Collis et al. (2004) recount the specific amendments to UK legislation governing statutory audit exemptions.

3

amendment to the UK Companies Act further relaxed the eligibility thresholds, culminating with private companies qualifying for the audit exemption in fiscal years ending after January 30, 2004 if their sales did not exceed £5.6m and total assets did not exceed £2.8m. Prior to this date, it had been permissible for private companies to avoid having an audit if their sales did not exceed £1m and their total assets did not exceed £1.4m. The shift in exemption eligibility thresholds in 2004 enables us to assemble a sample of companies that were affected by the regime switch.3

Fewer than 6,000

companies were affected by the rule change whereas there are more than one million private companies in the UK, implying that it is unlikely that the rule change had a major impact on the degree of competition within the private company audit market as a whole. Accordingly, we focus on the companies that were affected by the regime switch rather than the entire audit market. For each affected company, audits were mandatory in 2003 and voluntary afterward. Our analysis proceeds in two steps. First, we evaluate whether companies’ audit choices during the mandatory regime hinge on whether they genuinely want to be audited.

The motivation for this analysis is the uncertainty surrounding whether

companies not wanting to be audited can be compelled to undergo audits that are as stringent as those supplied to companies that would purchase audits even under a voluntary regime. Specifically, any companies that do not wish to be audited would likely minimize their costs under a mandatory regime. Such companies are likely to be “going through the motions” in of ively complying with the audit requirement, reducing the benefit of making audits mandatory. Consequently, we begin by examining companies’ audit choices during the final year of the mandatory audit regime. To identify whether companies genuinely wanted 3 We refer to the voluntary audit regime as starting in 2004 since none of the companies in our sample had year-ends falling between January 1 and January 29, 2004.

4

to be audited at this time, we observe their decisions on whether to remain audited in the following year when audits became voluntary.

We expect the companies not

wanting to be audited (as revealed in their subsequent decision to abandon an audit) under the mandatory regime were intent on minimizing their audit costs, rather than relying on auditing to provide strict external monitoring. Corroborating this intuition, we find that these companies were significantly less likely to hire Big Four auditors and they paid significantly lower audit fees during the last year of the mandatory regime. In contrast, the companies that wanted to be audited were significantly more likely to appoint high-quality auditors and they incurred significantly higher fees. These results suggest that the assurance benefits of mandatory audits are lower for companies that do not want to be audited relative to companies that would choose to be audited under a voluntary regime. This significantly weakens the main argument in favor of mandatory audits, which is based on the assumption that there are considerable assurance benefits when companies that do not want to be audited are forced to undergo an audit. Importantly, imposing audits prevents companies from signaling through their decision on whether to appoint an auditor. The second part of our analysis involves isolating whether the regime switch from mandatory to voluntary auditing yields signaling benefits by enabling outsiders to observe companies’ choices on whether to be audited. Melumad and Thoman (1990) provide a theoretical model in which companies requiring loans voluntarily choose whether to subject themselves to an audit. In their set-up, companies are composed of unobservable risk types and lenders draw rational inferences about the company’s type, conditional on its decision on whether to purchase an audit.

These authors demonstrate that auditors can be hired purely for their

signaling value. In their model, companies elect to hire auditors even when audits have no assurance value because lenders conclude that only the bad types of borrowers would choose not to be audited. In other words, a company would rather hire an

5

uninformative auditor and be pooled with the good types than avoid an audit, which divulges to lenders that they are a worse type. Costly auditing permits a separating equilibrium in which the good types of borrowers appoint auditors, unlike the bad types. It follows that the company’s decision to appoint (not appoint) an auditor sends a positive (negative) signal about its credit risk. Consistent with Melumad and Thoman’s (1990) intuition, critics of mandatory auditing argue that audits should be voluntary in order to facilitate the signaling of companies’ types. For example, Sunder (2003) writes: “All things being equal, investors can logically conclude that the firms that choose to have their financial reports audited by independent auditors have nothing to hide from the investors; that the managers of such firms are relatively more confident about the status, performance and prospects of their business; and that they deserve the trust of the investors. On the other hand, the investors may logically conclude that the firms which choose not to have their financial reports audited by independent auditors, even though they could have done so, are less deserving of the investors’ trust and money. When we make independent audit a statutory requirement, we shut the door on the ability of the better managers to distinguish themselves from the less competent managers in the eyes of the investors.” Despite this argument, we are unaware of any prior empirical evidence on the signaling role of voluntary audits.

We primarily contribute to extant research by assessing

whether voluntary audits have signaling value separate from their assurance value. Theory shows that companies are able to signal their types even when audits are mandatory since companies retain the discretion to hire either low-quality or highquality auditors (Titman and Trueman, 1986; Datar et al., 1991). Analytical research suggests that superior companies signal their types by spending more money on appointing high-quality auditors. The bad types of borrowers are more likely to appoint low-quality auditors when they are forced to purchase audits. The evidence in the first part of our analysis indicates that the strength of companies’ preferences for an audit

6

would have been partly revealed during the mandatory regime since the companies that were audited involuntarily were choosing non-Big Four auditors and were paying significantly lower audit fees. We therefore expect that outsiders can rationally infer a company’s likely preference for an audit even when audits are required for every company. The second part of our empirical analysis focuses on the company’s credit rating given that our sample consists of small private companies which rely mainly on external credit finance rather than outside equity. The credit rating is an important source of information for creditors deciding whether to lend and pricing the of their loans. Credit ratings are assigned to virtually all private companies in the UK by Qui Credit Assessment Limited. If the ratings agency is able to infer companies’ types even when all companies are required to be audited, we expect that the bad types of company that do not want to be audited would suffer lower ratings. In contrast, the good types that genuinely want to be audited would enjoy relatively high credit ratings even when they are unable to signal their types through their decision on whether to be audited. To gauge whether companies genuinely wanted to be audited during the mandatory regime, we examine their decisions on whether to be audited when audits subsequently become voluntary.

Consistent with our predictions, we find that

companies wanting to be audited received significantly higher ratings during the mandatory regime in contrast to companies that are audited involuntarily, which were penalized with lower ratings. These findings lend to the argument that the ratings agency was able to distinguish between companies’ types even when auditing was mandatory. This in turn implies that the switch to a voluntary regime would not necessarily provide an additional informative signal about companies’ types.

7

To isolate whether voluntary audits did in fact generate an incrementally informative signal, we examine the changes in credit ratings after the transition from mandatory to voluntary audits. When a company voluntarily chooses to continue being audited, we expect little or no change in audit assurance since the company is audited in both regimes. However, such a company conveys a positive signal when it chooses to be audited voluntarily since it is separated from the bad types that relinquish the audit. Therefore, to the extent that a voluntary audit communicates an incremental signal of the company’s superior type, we expect an increase in credit ratings for the companies that elect to remain audited.4 For a company that chooses to no longer engage an auditor, its decision signals not only that it is likely to be a bad type of borrower, but also sacrifices the assurance that was previously provided under the mandatory regime. For both reasons, we expect that credit ratings decline for companies that dispense with an audit when this becomes permissible. In regressions that control for changes in company characteristics and macroeconomic conditions, we provide strong, robust evidence that credit ratings rise (fall) for companies that continue (stop) being audited.

Specifically, credit ratings

increase by approximately two percentage points when companies choose to remain audited. We interpret this evidence as implying that these companies enjoy upgrades to their credit ratings because their decision to remain audited conveys a positive signal about their credit risk. Apparently, the level of audit assurance was stable for these companies during the transition from mandatory to voluntary audits since we find that their audit fees and choices of audit firm do not change following the regime switch. In

4 We do not assume that every unaudited company is a bad type or that every audited company is a good type. Rather, our predictions are based on the theoretical premise that, everything else constant, the good (bad) types have stronger (weaker) incentives to be audited (Titman and Trueman, 1986; Melumad and Thoman, 1990; Datar et al., 1991). Accordingly, the decision on whether to be audited sends a signal about the average type of company making that choice, although the signal is not necessarily indicative of every company’s true type.

8

the other direction, we find that credit ratings drop by approximately four percentage points when companies choose to stop being audited. For such companies, the financial statements become less credible because they are no longer audited and the decision to abandon the audit also conveys a negative signal about the company’s type. Our evidence on credit ratings is summarized as follows. During the mandatory regime, the companies that wanted to be audited receive credit ratings that were approximately nine points higher than those that did not.

After auditing becomes

voluntary, the companies that remain audited receive a further two point boost to their ratings while the unaudited companies suffer a four-point penalty. Consequently, the spread in credit ratings widens following the switch away from mandatory audits because the ratings agency is able to better distinguish between the good and bad types of borrowers.

During the voluntary regime, the audited companies receive credit

ratings that are fifteen points higher than the unaudited companies. This is comprised of the initial spread during the mandatory regime (i.e., nine points) plus the increase in the spread when auditing becomes voluntary (i.e., six points).

To provide some

perspective on the materiality of these results, the mean (median) credit rating over our sample period is 67 (69) points on a 100-point scale. Collectively, our evidence suggests that the switch to voluntary auditing had a first-order economic impact on credit ratings.5 In their comprehensive literature reviews, Francis (2004) and Watkins et al. (2004) outline the extensive prior theory and evidence that audit quality varies along several dimensions.

We complement this research by focusing on the more basic

5 In a sensitivity test, we dispel concern that auditor choice is spuriously responsible for our evidence on the importance of legal regime (mandatory versus voluntary) to credit ratings by ing that our core results hold when we run the regressions on non-Big Four (Big Four) clients separately.

9

question regarding the impact of allowing auditing to be optional rather than required. 6 To our knowledge, this is the first study to provide direct evidence on the signaling benefits of voluntary auditing. Blackwell et al. (1998) provide evidence that lenders charge lower interest rates to private companies whose financial statements are audited.7 Although their study indicates that there are benefits from being audited voluntarily, it does not disentangle whether such benefits stem from greater assurance or from signaling. Specifically, an audited company may enjoy a lower interest rate because its financial statements are more credible or because its decision to purchase an audit sends a positive signal about its credit risk. Thus, their finding that audited companies incur a lower cost of borrowing does not necessarily imply that voluntary audits yield signaling benefits. Likewise, there is evidence that lenders insist that companies supply audited financial statements to obtain loans (Leftwich, 1983; Allee and Yohn, 2009), although this could reflect either the assurance or signaling benefits of auditing. It is important to separately isolate the signaling effect because theory holds that⎯even if there is zero assurance value to an audit⎯companies would have incentives to appoint auditors voluntarily due to the positive signal that rational lenders would infer (Melumad and Thoman, 1990). In a major contribution to extant research, we provide the first direct evidence ing this theoretical prediction about the signaling value of voluntary audits. There is already fairly extensive evidence on the impact of voluntary audits, leading Francis et al. (2008: 1) to highlight that: “the most robust finding across these studies is that private [voluntary] audits are related to securing external debt financing.”

6

The demand for voluntary audits predates regulations requiring audits. Companies have been voluntarily appointing independent auditors since at least the 13th century (Watts and Zimmerman, 1983; Sunder, 2003), which implies that the private benefits of auditing often exceed their costs. Bentson (1969) reports that 82 percent of US public companies purchased audits shortly before the Securities Acts of 1933 and 1934, which introduced mandatory auditing of financial statements. Chow (1982) examines the voluntary audit decision made by large US public companies in 1926. Given the evolution in the institutions governing financial reporting, Barton and Waymire (2004) caution against generalizing results from earlier times to the present. Similarly, Coffee (1984) calls for research on contemporary regulation since so much has changed since the 1933 and 1934 Acts were ed. 7

10

Our evidence on signaling is also central to the debate about the relative merits of voluntary and mandatory audits since both types of audits can yield assurance benefits but only voluntary auditing facilitates signaling.

Our findings provide

empirical for the argument that the mandatory requirement suppresses the signal that is conveyed when companies are allowed to choose whether to be audited (Sunder, 2003). Moreover, our study indicates that companies not wanting to be audited were only ively complying with the audit requirement under the mandatory regime. These companies chose lower quality auditors and they were intent on reducing the costs of the audit when they were audited involuntarily. It would appear that there are limited assurance benefits arising from involuntary audits since it is difficult to force such companies to undergo stringent auditing. This in turn weakens the main argument in favor of mandatory auditing. Finally, it is important to acknowledge that our study cannot provide a definitive answer to policy-makers on whether audits should be voluntary or mandatory. One reason is that companies focus on their own private costs and benefits when considering whether to purchase an audit and we are not able to measure the external spillovers that accrue to external s from having the s audited for every company. Another reason is that our empirical analysis is restricted to small private companies and the results may not generalize to companies that are publicly traded. The remainder of this paper is organized as follows. Section 2 outlines prior theory and evidence in developing our testable predictions.

Section 3 covers our

research design, while Section 4 describes our empirical results. Section 5 concludes.

11

2. Hypotheses development The purpose of this paper is to shed light on the role of different legal requirements governing auditing. The rationale behind regulators imposing mandatory audits is that companies have insufficient private incentives to voluntarily provide reliable financial information.8 Reinforcing that mandatory auditing may be socially optimal, theory implies that information underproduction can arise because of positive externalities (e.g., Dye, 1990; ati and Pfleiderer, 2000) and financial statements constitute a public good that will be under-supplied in a free market (e.g., Gonedes and Dopuch, 1974; Beaver, 1998). On the other hand, it remains unclear whether private incentives for the demand and supply of audits are insufficient, particularly as private markets for other forms of certification services are ubiquitous in the economy (Jamal and Sunder, 2008). Moreover, voluntarily submitting to an audit enables companies to credibly signal their types whereas mandatory audits deprive investors of this important indicator (Sunder, 2003). The arguments in favor of regulation rather than the free market were influential in the UK, where private companies are required to make their financial statements publicly available and, until recently, those statements also had to be independently audited (Aranya, 1974; Dedman and Lennox, 2009).

The rationale for requiring all

companies to be audited hinges on the necessary condition that mandatory audits provide assurance benefits.

However, some companies may be going through the

motions when complying with this regulation, which would translate into mandatory audits delivering relatively low assurance.

8 Importantly, Dye (1990) and Leftwich (2004), among others, stress that advocates for mandatory financial reporting do not argue that firms will supply no information unless compelled or that managers’ incentives are irrelevant to reporting decisions. Indeed, the opportunity to reduce agency (Jensen and Meckling, 1976), information (Botosan, 1997), and liquidity (Welker, 1995) costs can convince managers to divulge high-quality information voluntarily.

12

We begin by testing whether companies that did not want to be audited⎯evident in their subsequent decision to avoid an audit when this becomes permissible⎯were only ively complying with the mandatory audit provision. Imposing audits precludes companies from signaling their types through the decision on whether to appoint an auditor, although they can still partly reveal their type with their decision on whether to choose a high-quality or low-quality auditor (Titman and Trueman, 1986; Datar et al., 1991). The bad types of borrowers that are reluctant to be audited will resist choosing high-quality auditors when they are forced to be audited. In contrast, the good types that genuinely want to be audited are more likely to choose high-quality audit firms when auditing is mandatory. The extensive empirical evidence that the Big Four firms supply higher quality audits than do the smaller auditors (e.g., Francis, 2004) extends to the U.K. (e.g., Lennox, 1999; Peel and Roberts, 2003; McMeeking et al., 2006; Clatworthy and Peel, 2007). Reflecting that the bad types of borrower are less willing to bear the costs of a high quality audit, their lower demand for audits translates into our first prediction (stated in the alternative form): H1:

During the mandatory audit regime, companies that do not want to be audited are less likely to choose a Big Four auditor compared with companies that do want to be audited.

Similar arguments apply to a company’s incentives on whether to pay for a highquality audit. A company that does not want to be audited will likely put pressure on the auditor to cut costs in order to reduce the audit fee. In contrast, a company that genuinely wants to be audited would be more willing to pay the high fee that ensues when an auditor supplies more effort on the assurance engagement. We therefore expect the companies that did not want to be audited were paying relatively low audit fees during the mandatory regime, which provides our second hypothesis: H2:

During the mandatory audit regime, companies that do not want to be audited pay lower audit fees compared with companies that do want to be audited.

13

Our next set of hypotheses concern the credit ratings assigned to private companies during the mandatory and voluntary regimes. Private companies provide an opportune setting for analyzing the assurance and signaling benefits of auditing given that their information structure is typically poor relative to public companies.

For

example, Brav (2009) holds that debt contracting is more sensitive to information for private companies than for their public counterparts.

Reinforcing that this testing

ground suits our inquiry, Fenn (2000) and Santos (2003) report that lenders demand higher yields on private companies’ debt to compensate for the worse information asymmetry that they suffer.9 This evidence squares with Graham et al.’s (2005) finding from a survey of chief financial officers that, compared with public companies, private companies are more inclined to manipulate earnings to preserve their credit ratings and to avoid violating bond covenants, rendering their financial statements less informative for the debt contracting process. It follows that the links between auditing and credit ratings will be stronger in private companies that are lesser known, increasing the power of our tests. Analyzing credit ratings maps into our research questions that focus on the benefits of auditing under the mandatory and voluntary regimes. Prior research shows that information risk affects credit ratings and that reliable ing numbers facilitate debt contracting (e.g., Watts, 1977; Smith and Warner, 1979; Leftwich, 1983; Francis et al., 2005; Yu, 2005). This is particularly relevant to UK private companies that rely heavily on loan financing (Brav, 2009).

In fact, prior research finds that companies relax

ing-based covenants by managing their earnings through ing changes (Sweeney, 1994) and discretionary accruals (DeFond and Jiambalvo, 1994). Moreover, 9 Similarly, Pagano et al. (1998) report that firms’ borrowing costs fall after going public, which they attribute to the greater ing transparency that ensues. Indeed, Givoly et al. (2009) report that public firms have more conservative earnings than private firms. Prior research implies that debt contracting is partly behind the demand for conservative financial reporting; e.g., Leftwich (1983) and Holthausen and Watts (2001).

14

recent research uses credit ratings to measure the perceived benefits of Big Four versus non-Big Four audits (e.g., Mansi et al., 2004), including in private companies (e.g., Fortin and Pittman, 2007).10 Outsiders can rationally infer the company’s type by observing its audit choices according to theory. Naturally, the low risk borrowers are eager to signal their types since they will be rewarded in the form of higher ratings for becoming better known. In analytical models, a separating equilibrium prevails in which companies with unfavorable private information choose low audit assurance while the good types choose high assurance. The bad types are deterred from mimicking the good types because it is too costly to appoint a high-quality auditor (e.g., Datar et al., 1991) or yields inadequate benefits (e.g., Titman and Trueman, 1986). Accordingly, we expect that the ratings agency is able to rationally infer whether a company genuinely wants to be audited even when every company is required to be audited. In short, we predict that credit ratings are lower for companies that do not want to be audited under the mandatory regime since the credit rating agency recognizes that these companies are more likely to be the bad types of borrowers: H3:

During the mandatory audit regime, companies that do not want to be audited receive lower credit ratings compared with companies that do want to be audited.

Auditors provide implicit insurance coverage to investors in the event of audit failure (Dye, 1993), although empirical research generally struggles with cleanly distinguishing between the information and insurance roles that auditing plays in securities pricing. However, auditor insurance protection is almost certainly trivial in our setting. St. Pierre and Anderson (1984) and Palmrose (1987) find that civil lawsuits against auditors infrequently involve US private firms despite that this country is a global outlier in of its high rate of auditor litigation (Francis, 2004). Similarly, Palmrose (1986) reports that the litigation risk embedded in audit fees is higher for public companies than private companies. In the UK, legal standards for suing auditors for issuing an unqualified opinion on materially deficient financial statements are stricter (Seetharaman et al., 2002; La Porta et al., 2006); i.e., relative to the litigious environment in US, auditor discipline is more lenient in the UK where it is more difficult for investors to sue the auditor to recover losses. In fact, auditors only owe a duty of care to shareholders, not creditors, under UK case law. Reinforcing the minor role of litigation there, the UK prohibits class-action lawsuits.

10

15

Finally, we analyze whether credit ratings impound that requiring audits ruins the signaling value inherent in companies electing to subject their financial statements to an external audit. Our identification strategy exploits the new UK legislation to provide insight on whether companies enjoy higher credit ratings when they more fully reveal their types by choosing to appoint an auditor under the voluntary regime.

In the

absence of an audit requirement, a company can incur the cost of engaging an auditor to ensure that lenders infer that it is more likely to be a good type.

More formally,

Melamud and Thoman’s (1990) theory implies that the information conveyed with the decision to hire an auditor vanishes when auditing becomes mandatory. A mandatory audit regime brings uncertainty by preventing lenders from learning about a company’s type by observing its strategic decision on whether to have an audit (Sunder, 2003); i.e., requiring audits suppresses this form of information. In estimating the signaling value of voluntary audits, the benefits of audit assurance are held constant for companies that are audited in both the mandatory and voluntary regimes. On the other hand, there is a role for signaling under the voluntary regime as a company that continues to be audited is able to signal that it is more likely to be a good type. In this case, remaining audited under the voluntary regime transmits positive news evident in higher ratings. H4:

Credit ratings increase for companies that switch from mandatory audits to voluntary audits.

In contrast, companies that choose to stop being audited under the voluntary regime forego any assurance benefits that existed when they were involuntarily audited and their decision also conveys a negative signal because outsiders are able to more clearly distinguish between them and the companies that want to be audited. These dynamics are behind our final hypothesis:

16

H5:

Credit ratings fall for companies that switch from mandatory audits to no audits.

Collectively, we expect that credit ratings are higher for companies that want to be audited than for those that do not, and the spread in credit ratings between the two types of company expands when audits become voluntary since this enables companies to more fully reveal their types. Although requiring audits dilutes their signaling value, we expect that companies still partly divulge their type through audit fees and auditor choice when audits are mandatory.

3. Research Design 3.1 Final Year of the Mandatory Regime To test the first three hypotheses, H1 to H3, our analysis initially focuses on the final year of the mandatory regime. Companies that do not want to be audited may be simply going through the motions when audits are required, which would be evident in these companies choosing lower quality audit firms (H1) and paying lower audit fees (H2) during the mandatory regime and then avoiding an audit during the voluntary regime. In contrast, the companies that genuinely want to be audited are more likely to be perceived as low-risk borrowers. If the ratings agency is able to discriminate between the two types of companies even when they are both subject to an audit, we would expect that the companies that want to be audited enjoy higher credit ratings (H3). We test these three predictions by estimating the following cross-sectional models for the final year of the mandatory audit regime (2003): BIG4it = β0 + β1 VOL_AUDITit + β2 Xit + β3 INDUSTRYi + vit

(1)

LAFit = α0 + α1 VOL_AUDITit + α2 Xit + α3 INDUSTRYi + uit

(2)

RATING it = δ0 + δ1 VOL_AUDIT it + δ2 Xit + δ3 INDUSTRYi + wit

(3)

17

Where: BIG4it = one if company i chooses a Big Four audit firm in 2003, and zero if company i chooses a non-Big Four audit firm in 2003. LAFit = the log of company i’s audit fees in 2003. RATINGit = company i’s credit rating in 2003 (the rating is scored by Qui Credit Assessement Ltd. on a scale ranging from 0 to 100). VOL_AUDITit = one if company i is voluntarily audited in 2004, and zero if company i is unaudited in 2004. Xit = a vector of time-varying control variables for company i (t = 2003). INDUSTRYi = a vector of industry dummy variables for company i. The coefficients on the VOL_AUDITit variable will be significantly positive in eqs. (1) to (3) under H1 to H3.

3.2 Transition from the Mandatory to the Voluntary Regime Next, we analyze the changes in credit ratings after the regulatory regime switches from mandatory to voluntary audits. We expect that credit ratings will rise for companies that remain audited when this becomes voluntary in 2004 since electing to have an audit would serve as an incremental signal of the company’s good type (i.e., lower credit risk). In comparison, abandoning an audit is both a signal that the company is a likely to be a bad type of borrower and it removes the assurance value associated with having audited financial statements.

Both effects are expected to contribute towards lower credit

ratings. Accordingly, we examine the predictions in H4 and H5 by estimating a model that explains how credit ratings change from the final year of the mandatory regime

18

(2003) to the initial year of the voluntary regime (2004). Re-writing the credit ratings model in eq. (3) by taking changes gives the following:

ΔRATINGit = µ0 + µ1 VOL_AUDIT it + µ2 ΔXit + wit

(4)

where

ΔRATINGit = the change in company i’s credit rating from the last year of the mandatory audit regime (2003) to the first year of the voluntary audit regime (2004). VOL_AUDITit = one if company i is voluntarily audited in 2004, and zero if company i is not audited in 2004.11 We expect that credit ratings increase for the companies that continue to be audited (H4) and fall for the companies that switch from mandatory audits to no audits (H5). tly, these two hypotheses generate the prediction that µ1 > 0. To provide a separate test of H4, we estimate eq. (4) for the sub-sample of companies that continue to be audited (i.e., VOL_AUDITit = 1):

ΔRATINGit = µ0 + µ1 + µ2 ΔXit + wit

(4a)

Under H4, credit ratings increase for a company that remains audited because this decision conveys a positive signal about its type (i.e., µ0 + µ1 > 0). Similarly, we test H5 by estimating eq. (4) for the sub-sample of companies that no longer undergo an audit (i.e., VOL_AUDITit = 0).

ΔRATINGit = µ0 + µ2 ΔXit + wit

(4b)

Under H5, credit ratings decrease for the companies that dispense with an audit (i.e., µ0 < 0) because this decision both results in a loss of assurance and conveys a negative signal about the company’s credit risk. 11 The VOL_AUDIT variable reflects both the decision to have an audit in 2004 (eq. 3) and the it change in audit status from 2003 to 2004 (eq. 4) since every company is audited during the final year of the mandatory regime. It is therefore equivalent to a change variable in eq. (4).

19

3.3 Control Variables Our set of control variables follows that used in recent research on credit risk, auditor choice, and audit fees (e.g., Bharath et al., 2008; Fortin and Pittman, 2007): AGE it = company i’s age in year t. LTSit = log of total sales. LTAit = log of total assets. INTCOVit = interest expense divided by earnings before interest and taxation. The INTCOVit ratio is capped at an upper bound of 2.00 to handle outliers and we assign a value of 2.00 to the INTCOVit ratio if earnings are non-positive. LIQUIDITYit = (current assets – inventory) divided by current liabilities. LEVERAGEit = total liabilities divided by total assets. We expect that insolvency risk is higher for companies that are younger (AGEit), smaller (LTSit and LTAit), have lower interest coverage (INTCOVit), lower liquidity (LIQUIDITYit), and more liabilities (LEVERAGEit).

Besides explaining credit ratings,

prior research implies that these independent variables ⎯ particularly company size ⎯ are important determinants of the company’s choice of audit firm and audit fees.12

Some audit fee studies include a control for profitability or a dummy variable for losses. However, in our study, profitability is very highly correlated with the interest coverage variable which measures the ratio of the interest expense to earnings. Given this high correlation we do not include the profitability and interest coverage variables in the same regressions. However, in robustness tests, we find very similar results if we replace the interest coverage variable with alternative measures of profitability.

12

20

3.4 Sample Formation and Description The UK has long diverged from the US by stipulating that private companies have their financial statements audited.13 However, in an effort to reduce the burden of regulation on small private companies, the UK granted an audit exemption in 1994. It became permissible after 1994 for private companies to avoid having an audit if their sales did not exceed £1m and their total assets did not exceed £1.4m. A subsequent amendment to the Companies Act relaxed these size thresholds, allowing more private companies to qualify for the audit exemption. Specifically, companies with fiscal years ending after January 30, 2004 were allowed to avoid an audit if their sales did not exceed £5.6m and total assets did not exceed £2.8m. This shift in exemption eligibility thresholds enables us to assemble a sample of companies that were affected by the regime switch in 2004. We compile the sample from the Financial Analysis Made Easy (FAME) database. All public companies must be audited regardless of their size, so we require that each sample company is private. For companies belonging to a group, auditor hiring decisions are routinely made by the ultimate owner rather than at the company level, so we impose the restriction that the company is independent; i.e., it does not belong to a corporate group. To ensure that each private company was required to undergo an audit prior to January 30, 2004, we sample companies for which sales ≥ £1m and total assets ≥ £1.4m in 2003.

Next, we confine the sample to companies that

qualified for the audit exemption after January 30, 2004 (i.e., sales ≤ £5.6m and total assets ≤ £2.8m). Certain types of regulated company⎯insurance companies, investment More generally, this setting is unique in other ways, including that both private and public companies must publicly file annual financial statements that follow the same ing standards in order to comply with UK disclosure regulations. Similarly, both types of companies are subject to the same tax laws in the UK.

13

21

advisors, mortgage arrangers, trade unions, and employers associations⎯are required to have an audit even if they fall within these size thresholds, so we exclude these companies from the sample.

Finally, we require that financial statement data are

available for both the year prior to January 30, 2004 and the next year. After applying these data screens, we are left with 5,139 unique companies. By design, each company in the sample was required to have an audit in 2003, but not in 2004. There are two observations per company with the first pertaining to the final year of the mandatory audit regime (2003) and the second to the initial year of the voluntary regime (2004). 3,440 (67%) of the 5,139 companies in the sample remain audited in the first year of the voluntary regime while 1,699 companies (33%) choose to become unaudited once this option becomes available.

3.5 Descriptive Statistics The industry composition of the sample companies is shown in Table 1. There are 1,414 companies (26.3%) that operate in the business services sector and 1,221 (23%) companies in the wholesale and retail trade sector.

Other industries that are well

represented include construction (748 companies) and manufacturing (719 companies). Our research design controls for differences between industries by specifying dummy variables for the eight industry sectors in Table 1 that have more than twelve companies (the remaining three industry sectors are captured in the regression intercept). [INSERT TABLE 1 NEAR HERE] The credit rating scores issued by Qui Credit Ltd. are provided in the FAME database. Ratings are assigned on a numerical scale between 0 and 100 that quantifies the agency’s assessment of the likelihood of corporate failure within the next 12 months.

22

The RATINGit variable corresponds to the company’s credit score with higher ratings representing a lower perceived risk of financial failure. Doumpos and Pasiouris (2005) provide evidence that Qui’s credit ratings are accurate indicators of default risk. In our own analysis, we find that the credit scores are strongly associated with staple ing variables used in predicting insolvency risk (e.g., interest coverage). Table 2 presents descriptive statistics for 2003 (the final year of the mandatory regime) and 2004 (the first year of the voluntary regime).

The motivation for this

analysis is to investigate whether there are important changes in the macroeconomic environment or other time-varying factors that could confound the comparison of the two regimes. The first row of Table 2 shows that the mean credit rating increases from 64.51 in 2003 to 68.37 in 2004. This ratings improvement, which is highly significant (tstatistic = 9.32), occurs despite the potential loss of assurance that follows when companies start to abandon audits in 2004. The ratings increase is consistent with a general improvement in the economic environment between 2003 and 2004. Further, we find a significant improvement in liquidity between 2003 and 2004 (t-statistic = 2.73), although the changes in interest coverage, company size, and leverage are not significant. [INSERT TABLE 2 NEAR HERE] Given the evidence in Table 2, we control for the general improvement in credit ratings that affects our entire sample between 2003 and 2004 by subtracting the sample mean values of credit ratings in each year. For example, we measure the deviation between company i‘s credit rating in 2003 and the mean rating given to every other company in 2003 (RATINGi2003 - RATING.2003). We similarly measure the cross-sectional variation in credit ratings during 2004 using the variable RATINGi2004 - RATING.2004.

23

Finally, we calculate the change in credit ratings from 2003 to 2004 using the ratings that are purged of the year effects. This change in credit ratings variable is (RATINGi2004 RATING.2004) – (RATINGi2003 – RATING.2003), which we label Δ(RATINGit – RATING.t). Similarly, we purge all the control variables of any yearly effects following the same approach; i.e., Δ(Xit – X .t). Thus, we modify the models of credit ratings changes in eqs. (4), (4a) and (4b), which now become:

Δ(RATINGit – RATING.t) = µ0 + µ1 VOL_AUDITit + µ2 Δ(Xit – X .t) + wit

(4’)

Δ(RATINGit – RATING.t) = µ0 + µ1 + µ2 Δ(Xit – X .t) + wit

(4a’)

Δ(RATINGit – RATING.t) = µ0 + µ2 Δ(Xit – X .t) + wit

(4b’)

4. Results 4.1 Univariate Evidence In an initial at our research questions, we report in A of Table 3 the mean values of each variable during the mandatory regime. The BIG4it frequency is 8% in the sub-sample of 3,440 companies that wanted to be audited, which is evident in their decision to be voluntarily audited in the subsequent year (VOL_AUDIT it = 1). The BIG4it frequency is significantly lower at 2% for the 1,699 companies that did not want to be audited (VOL_AUDIT it = 0), consistent with the prediction in H1 that the companies genuinely eager to be audited were significantly more likely to appoint Big Four audit firms (t-statistic = -8.16). [INSERT TABLE 3 NEAR HERE] Similarly, the mean audit fee is £5,680 for the companies that wanted to be audited compared with £4,270 for the companies that did not. These audit fees are naturally lower than in prior studies since our sample consists of small private

24

companies (the average company has assets hardly exceeding £1 million according to Table 2). More importantly for our purposes, we find that audit fees are significantly lower for the companies that did not want to be audited (t-statistic = -11.87), ing the prediction in H2. The highly significant difference in the log of audit fees (LAFit) between the two groups of companies (t-statistic = -13.32) reinforces this result. Consistent with H3, it appears that the credit ratings agency perceived that the companies wanting to be audited were significantly lower risks.

The mean rating

assigned to these companies was 67.48 compared with 58.50 for the companies that did not want to be audited; this nine-point difference is highly significant (t-statistic = -13.64). While these univariate results are consistent with all three hypotheses, it is important to caution that they do not control for the observable differences between the two types of company. For example, according to A of Table 2, the companies that wanted to be audited are significantly older (t-statistic = -5.11) and larger (t-statistics = -4.21, -12.80). In addition, these companies had worse interest coverage (t-statistic = -3.50), although they also had greater liquidity (t-statistic = -1.87) and lower leverage (tstatistic = 3.63). B of Table 3 reports the mean values during the first year of the voluntary audit regime.

Interestingly, we find that the spread in credit ratings between the

companies that choose to be audited and those that do not is larger in 2004. The mean rating to the voluntarily-audited companies is 73.38 compared with 58.25 for the companies that chose not to have their financial statements audited. The difference in the ratings assigned to these two types of company is more than fifteen points during the voluntary regime (73.38 minus 58.25) compared with only nine points in the

25

previous year (67.48 minus 58.50). This wider spread is our first piece of evidence that the regime switch affected the ratings issued to the two groups of companies. It is important to consider the alternative explanation that the changes in credit ratings were different for the two sets of companies because they experienced significant differences in their performance between 2003 and 2004. The results in s C and D of Table 3, however, clearly dispel this concern. In these s, we compare the mean values of the ratings and control variables between 2003 and 2004, after purging the variables of their yearly mean values (i.e., these s report the mean values of RATINGit – RATING.t and Xit – X .t). C provides no evidence of significant changes in performance or operating characteristics between 2003 and 2004 for the 1,699 companies that did not want to be audited. In fact, the mean values are statistically insignificant between these two years for every control variable. In contrast, C shows that credit ratings fell dramatically between 2003 and 2004 for the companies that avoid an audit. Their mean credit rating was 6.01 points below the mean in 2003 whereas it was 10.13 points below the mean in 2004. Therefore, consistent with H5, these companies suffered a relative downgrade of just over four points in their credit ratings. In D, we report descriptive statistics for 2003 and 2004, focusing on the 3,440 companies that remain audited following the regime switch. The first two rows reveal that the audit fees paid by these companies did not change significantly and their choices of audit firm also did not change. Therefore, there was no apparent change in audit assurance between 2003 and 2004 for the companies that were audited in both years. Nevertheless, we expect that their credit ratings would improve over time since voluntarily submitting to an audit signals their superior types.

ing this

argument, D shows that credit ratings are 5.00 points above the mean in 2004, but

26

only 2.97 points above the mean in 2003 for these companies. In short, consistent with H4, the companies that voluntarily continued with an audit enjoyed a two-point ratings upgrade. Again, we find no evidence that changes in company characteristics explain this change in credit ratings. D shows that these companies exhibit no significant changes in interest coverage, size, liquidity, or leverage between 2003 and 2004. Collectively, the results in s C and D strongly suggest that the changes in credit ratings between 2003 and 2004 are attributable to the regime switch, rather than changes in company characteristics.

4.2 Auditor choice and audit fees during the mandatory audit regime Companies that do not want to be audited may be simply going through the motions when they are forced to be audited during the mandatory regime. This would be evident in these companies attempting to minimize audit costs during the mandatory regime and then avoiding an audit during the voluntary regime. In contrast, companies genuinely eager to improve financial reporting credibility would accept incurring the higher fees that accompany a higher-quality audit.

We provide evidence on the

predictions in H1 and H2 by estimating models on the determinants of auditor choice (eq. 1) and audit fees (eq. 2) during the final year of the mandatory audit regime.14 The results for the auditor choice model are shown in Col. (1) of Table 4. Consistent with H1, the VOL_AUDITit coefficient is positive and highly significant (t-stat. = 5.94). That is, the companies that wanted to be audited were more likely to appoint Big Four audit firms during the mandatory regime compared with the companies that Although prior research finds that Big Four auditors charge higher fees, the pair-wise correlation between auditor choice and audit fees is only 0.14 in our sample, suggesting that these variables reflect different underlying constructs.

14

27

were ively complying with the audit requirement.

The results for the audit fee

model are shown in Col. (2) of Table 4. The VOL_AUDITit coefficient is again positive and highly significant (t-stat. = 7.20). This s the prediction in H2 that companies not wanting to be audited resort to paying lower audit fees. In contrast, the companies that wanted to be audited had a greater demand for assurance services evident in their higher audit fees. Overall, our findings are consistent with the theoretical prediction that companies’ types can be partially signaled through their audit choices even when auditing is mandatory (Titman and Trueman, 1986; Datar et al., 1991). However, it remains to be resolved whether the switch to voluntary auditing helped to more fully distinguish companies’ types. [INSERT TABLE 4 NEAR HERE] The results for the control variables are consistent with prior studies on the determinants of auditor choice and audit fees. The auditor choice model shows that Big Four clients tend to be larger, while they also have higher liquidity and greater interest coverage. Larger and older companies also pay significantly higher audit fees. The INTCOVit variable has a significantly positive coefficient in the audit fee model, implying that fees are higher when a company’s earnings are low relative to its interest expense. This could reflect that audit fees are higher when companies are in financial distress to compensate for increased audit risk.15

The R2 in the audit fee model is only 21%, which is lower than in prior studies because the companies in our sample are nearly homogeneous in of their size. Most of the R2 in prior audit fee studies comes from the cross-sectional variation in company size which is low in our study because every company must meet the size thresholds for mandatory audits in 2003 and voluntary audits in 2004. Moreover, it is important to note that the R2 statistics are not comparable across studies that examine very different samples (Gu, 2007). 15

28

4.3 Credit ratings during the mandatory and voluntary audit regimes We report the estimation results for the model of credit ratings during the mandatory regime (eq. 3) in Col. (1) of Table 5.

In this regression, the VOL_AUDITit coefficient is

positive and highly significant (t-stat. = 13.41), consistent with the prediction in H3 that the credit rating agency perceives that the companies that want to be audited have lower credit risk. The coefficient estimate on the VOL_AUDITit variable is 7.68 indicating that credit ratings are nearly eight points higher for the companies that wanted to be audited. The magnitude of this multivariate coefficient is very similar to the univariate result reported in A of Table 3, where the spread in ratings between the two groups of companies amounted to nine points. The similarity in results between the univariate and multivariate tests is comforting as it suggests that the results are unlikely to be affected by extraneous independent variables. Overall, we conclude that there are systematic differences in the credit ratings of voluntary and no audit companies even when auditing is mandatory. In particular, the credit rating agency considers the companies that wanted to be audited to be the superior types of borrowers in that they receive higher credit ratings. This is despite the fact that such companies are unable to fully reveal their types at this time through the signaling mechanism of choosing to be voluntarily audited. [INSERT TABLE 5 NEAR HERE] Next, we report the results for the model of credit ratings during the voluntary audit regime (Col. (2) of Table 5). In this regression, the coefficient on the VOL_AUDITit variable is 12.94 and it is again significantly greater than zero (t-statistic = 29.55). This finding implies that voluntarily-audited companies received credit ratings that are nearly thirteen points higher compared with the no audit case. Again, this magnitude

29

reinforces the univariate result ( B, Table 3), where it was found that credit ratings are fifteen points higher for the companies that are voluntarily audited. Further, Table 5 confirms that there is an increase in the credit ratings spread between the two types of company after the change in the regulation. This is consistent with the t prediction in H4 and H5, that credit ratings increase (decrease) for companies that want (do not want) to be audited between 2003 and 2004. We report separate tests of these two hypotheses in the next section. The results for the control variables are generally consistent with prior research that examines the determinants of private company credit ratings; e.g., Fortin and Pittman (2007). In particular, companies have better credit ratings when they are older (AGEit) and larger (LTSit, LTAit). In addition, companies with lower interest coverage (INTCOVit), higher leverage (LEVERAGEit), and lower liquidity (LIQUIDITYit) attract worse ratings.

4.4 Changes in credit ratings after auditing becomes voluntary The first model in Table 6 presents estimation results for eq. (4’) where the dependent variable (Δ(RATINGit – RATING.t)) captures the change in credit ratings as companies adjust to the voluntary regime (2004) from the mandatory regime (2003). As explained in Section 3.5, we subtract the yearly sample means for each variable in order to control for time-varying factors such as the macroeconomic environment. Table 6 does not include a variable for the change in company age since every company in the sample grows older by one year from 2003 to 2004; i.e., the change in age is the same for every company.

30

In Model 1, the VOL_AUDITit variable gauges the change in credit rating between 2003 and 2004 for companies that remain audited compared with companies that are no longer audited.

The coefficient on this variable is positive and highly

significant (t-statistic = 10.61), lending to the t prediction in H4 and H5 that there is a relative decline in credit ratings for companies that switch from mandatory audits to no audits, compared with companies that continue to be audited.

Reflecting

its economic materiality, the VOL_AUDITit coefficient is 5.84 in Col. (1), which implies that credit ratings increase (decrease) by nearly six points when companies continue being audited rather than avoid the audit. This magnitude is very similar to the increase in the credit ratings spread estimated in Table 5, where the difference in the VOL_AUDITit coefficients is 5.26 (= 12.94 – 7.68). [INSERT TABLE 6 NEAR HERE] Model 2 in Table 6 presents estimation results for the 3,440 companies that switch from mandatory audits to voluntary audits (eq. 4a’). The intercept in this model is estimated to be 1.93 and it is significantly greater than zero (t-statistic = 5.99). This implies that, between 2003 and 2004, credit ratings increase by nearly two points for the companies that continue to be audited following the regime switch. (This finding can also be calculated directly from model 1 since 1.93 = -3.91 + 5.84.) Consistent with H4, we conclude that companies transmit a positive signal about their credit risk by voluntarily continuing with an audit. This finding is important because it s the argument that voluntary auditing permits a signaling role that is absent when audits are required by law. To our knowledge, this study is the first to isolate the signaling effect stemming from allowing audits to be voluntary.

31

In Model 3, we report the changes in credit ratings for the 1,699 companies that choose not to be audited when this becomes permissible in 2004 (eq. 4b’). Consistent with H5, the companies that switch from mandatory audits to no audits experience an average downgrade of 3.91 points to their credit ratings. This drop in credit ratings is significantly different from zero (t-statistic = -8.81), corroborating the intuition that there is both a negative signal and a loss of assurance for the companies that cease to be audited. It is beyond the scope of this study to estimate the fraction of the 3.91 fall in credit ratings that reflects the negative signal versus the loss of audit assurance since these effects coincide.

4.5 Additional Analysis on Big Four and Non-Big Four clients In 2003, 4,835 sample companies were audited by non-Big Four firms compared with only 304 companies that were audited by the Big Four. These descriptive statistics reflect that all of the companies in our sample are very small, translating into relatively few companies needing to approach the Big Four firms to undertake their audits. In 2004, the number of Big Four audits falls by 18.4% from 304 to 248, while the number of non-Big Four audits drops 34.0% from 4,835 to 3,192. Accordingly, consistent with the evidence in of H1, the regime switch had a bigger impact on the number of audits conducted by the non-Big Four firms than those by the Big Four. Given that Big Four audits are relatively scarce in our sample, we repeat the tests of H2 to H5 by re-estimating the audit fee and credit ratings models after excluding the companies that were audited by Big Four firms in 2003. Consistent with H2, the audit fee model continues to show that the clients that did not want to be audited during the mandatory regime paid significantly lower audit fees. The models in Table 5 reveal that

32

these companies also received significantly lower credit ratings compared with those that wanted to be audited, which s H3.

Further, the credit ratings spread

between the two types of company increased from nearly eight points to thirteen points between 2003 and 2004. Finally, the ratings change models (Table 6) reveal that the nonBig Four clients enjoyed a statistically significant two point increase in their credit ratings if they continued being audited in 2004 (H4), while the non-Big Four clients that stopped being audited suffered a significant four point drop (H5). Predictably, the statistical significance and the economic magnitude of these results are very similar to our main tabulated specifications given that the vast majority of audits in our sample are performed by the non-Big Four firms. We also re-estimate the audit fee and credit ratings models for the sub-sample of 304 companies that were audited by Big Four firms in 2003.

In Table 4, the

VOL_AUDITit coefficient is just 0.02 (compared with 0.13 in the full sample) and it is not statistically different from zero. This is perhaps unsurprising given that the companies that chose Big Four audits had already revealed their preference for high-quality auditing. Thus, our results for H2 are significant only for the companies that hired nonBig Four firms during the mandatory regime. The credit ratings models in Table 5 continue to provide for H3. Specifically, the clients that wanted to be audited received credit ratings that were nearly seven points higher during the mandatory regime (t-statistic = 1.61) and fourteen points higher during the voluntary regime (tstatistic = 5.21). Finally, the models in Table 6 reveal that credit ratings increased for the 268 Big Four clients that chose to remain audited in 2004 (t-statistic = 4.08), while there was an insignificant drop in credit ratings for the 36 Big Four clients that were no longer audited (t-statistic = -0.48).

33

5. Conclusions The paucity of evidence on the implications of forced versus voluntary auditing motivates our research on economic outcomes surrounding legislation that rescinds mandatory audits for small private companies in the UK. arguments.

We analyze two main

First, we expect the companies that did not want to be audited were

privately contracting for a relatively low level of audit assurance during the mandatory regime. Second, we argue that there is a signaling benefit from allowing audits to be voluntary since the company’s decision on whether to be audited conveys valuable information to outsiders about its type. To test these predictions, we exploit a natural experiment in which audits were required for companies with fiscal years ending before January 30, 2004 and voluntary afterward. We examine both the audit choices that these companies made when audits were mandatory and the consequences of the change in the regulatory regime. Two major insights stem from our empirical analysis. First, the companies that did not want to be audited were significantly less likely to appoint Big Four audit firms and they paid significantly lower fees during the mandatory regime relative to the companies that wanted to be audited. These results suggest that such companies were ively complying with the audit requirement and they were subject to less strict monitoring than the companies that genuinely wanted to be audited. The main policy rationale for mandatory auditing is that external stakeholders obtain significant assurance benefits when companies that would not voluntarily choose an audit are forced into it. Our results suggest that these benefits are likely to be modest since the companies that do not want to be audited privately contract for low levels of audit assurance when audits are legally required.

34

Our second major finding is that the move away from mandatory auditing engenders an important role for signaling, in which the low-risk companies conveyed their favorable borrowing characteristics by continuing with an audit. Given recent research that auditors play a major role as information intermediaries in debt contracting, credit ratings suit our analysis of the signaling impact of the change in audit regime on the rating agency’s perception of borrower risk. The companies that remain audited enjoyed significantly higher credit ratings following the regime change even though these companies were audited in both 2003 and 2004 and the assurance value of their audits apparently did not change. We attribute this upgrade in credit ratings to the positive signal that voluntarily submitting to an audit sends to outsiders. Importantly, the decision to be audited voluntarily conveys information that is incremental to the signal that would occur in a mandatory regime, where companies turn to alternative mechanisms ⎯ namely, appointing Big Four auditors and paying higher audit fees ⎯ which can partly reveal their types.

This s the main argument in favor of

voluntary auditing, which is that imposing audits prevents companies from more fully revealing their types.

35

References ati, A. and P. Pfeiderer. 2000. “Forcing firms to talk: Financial disclosure regulation and externalities.” Review of Financial Studies 13: 479-515. Allee, K. D. and T. L. Yohn. 2009. “The demand for financial statements in an unregulated environment: An examination of the production and use of financial statements by privately held small businesses.” The ing Review 84: 1-25. Aranya, N., 1974. “The influence of pressure groups on financial statements in Britain.” Abacus June, 3–12. Reprinted in Parker, R.H. and T.A. Lee, eds. “The Evolution of Company Financial Reporting” (Thomas Nelson and Sons, Ltd., 1979), 265274. Barton, J. and G. Waymire. 2004. “Investor protection under unregulated financial reporting.” Journal of ing and Economics 38: 65-116. Beaver, W. H. 1998. Financial Reporting: An ing Revolution. Prentice-Hall, Engelwood Cliffs, NJ. Bentson, G. 1969. “The value of SEC’s ing disclosure requirements.” ing Review 54: 515-532. Bharath, S. T., J. Sunder, and S. V. Sunder. 2008. contracting.” The ing Review 83: 1-28.

The

“ing quality and debt

Blackwell, D., T. Noland, and D. Winters. 1998. “The value of auditor assurance: Evidence from loan pricing.” Journal of ing Research 36: 57-70. Botosan, C. A. 1997. “Disclosure level and the cost of equity capital.” The ing Review 72: 323–349. Brav, O. 2009. “How does access to the public capital market affect firms’ capital structure?” Journal of Finance 64: 263-308. Chow, C. W. 1982. “The demand for external auditing: Size, debt and ownership influences.” The ing Review 57: 272-291. Clatworthy, M. A. and M. J. Peel. 2007. “The effect of corporate status on external audit fees: Evidence from the UK.” Journal of Business Finance & ing 34: 169-201. Coffee, J. 1984. “Market failure and the economic case for a mandatory disclosure system.” Virginia Law Review 70, 717-753.

36

Collis, J., R. Jarvis, and L. Skerratt. 2004. “The demand for the audit in small companies in the UK.” ing and Business Research 34: 87-100. Datar, S. M., G. A. Feltham, and J. S. Hughes. 1991. “The role of audits and audit quality in valuing new issues.” Journal of ing and Economics 14: 3-49. Dedman, E., and C. Lennox. 2009. “Perceived competition, profitability, and the withholding of information about sales and the cost of sales.” Journal of ing and Economics, forthcoming. DeFond, M. and J. Jiambalvo. 1994. “Debt covenant violation and manipulation of accruals.” Journal of ing and Economics 17: 145-176. Doumpos, M. and F. Pasiouris. 2005. “Developing and testing models for replicating credit ratings: A multicriteria approach.” Computational Economics 25: 327–341. Dye, R. A. 1990. “Mandatory versus voluntary disclosures: The case of financial and real externalities.” The ing Review 65: 1-24. Dye, R. A. 1993. “Auditing standards, legal liability, and auditor wealth.” Journal of Political Economy 101: 887-914. Fenn, G. W. 2000. “Speed of issuance and the adequacy of disclosure in the 144A highyield debt market.” Journal of Financial Economics 56: 383-405. Fishman, M. and K. Hagerty. 1989. “Disclosure decisions by firms and the competition for price efficiency.” Journal of Finance 44: 633-646. Fortin, S. and J. A. Pittman. 2007. “The role of auditor choice in debt pricing in private firms.” Contemporary ing Research 24: 859-896. Francis, J. R. 2004. “What do we know about audit quality?” The British ing Review 36: 345-368. Francis, J. R., I. K. Khurana, X. Martin, and R. Pereira. 2008. “Private audits in developing and developed countries.” Working paper: University of MissouriColumbia. Francis, J., R. LaFond, P. Olsson, and K. Schipper. 2005. “The market pricing of accruals quality.” Journal of ing and Economics 39: 295-327.

37

Givoly, D., C. Hayn, and S. Katz. 2009. “Does public ownership of equity improve earnings quality?” Working paper: Penn State University, UCLA, and Harvard University. Gonedes, N. J. and N. Dopuch. 1974. “Capital market equilibrium, information production, selecting ing techniques: Theoretical framework and review of empirical work.” Journal of ing Research 12: 48-129. Graham, J. R., C. R. Harvey, and S. Rajgopal. 2005. “The economic implications of corporate financial reporting.” Journal of ing and Economics 40: 3-73. Gu, Z. 2007. “Across-sample incomparability of R2s and additional evidence on value relevance changes over time.” Journal of Business Finance and ing 34: 1073– 1098. Holthausen, R. and R. Watts. 2001. “The relevance of the value-relevance literature for ing standard setting.” Journal of ing and Economics 31: 3-75. Jamal, K. and S. Sunder. 2008. “Are regulatory mandate and independence necessary for audit quality?” Working paper, University of Alberta and Yale University. Jensen, M. C. and W. H. Meckling. 1976. “Theory of the firm: Managerial behavior, agency costs, and ownership structure.” Journal of Financial Economics 3: 305-360. Lambert, R., C. Leuz, and R. Verrecchia. 2007. “ing information, disclosure, and the cost of capital.” Journal of ing Research 45: 385-420. La Porta, R., F. Lopez-de-Silanes, and A. Shleifer. 2006. “What works in securities laws?” Journal of Finance 61: 1-32. Leftwich, R. 1983. “ing information and private markets: Evidence from private lending agreements.” The ing Review 58: 23-42. Leftwich, R. 2004. “Discussion of: Investor protection under unregulated financial reporting.” Journal of ing and Economics 38: 117-128. Lennox, C. 1999. “Are large auditors more accurate than small auditors?” ing and Business Research 29: 217-227. Leuz, C. and P. Wysocki. 2008. “Economic consequences of financial reporting and disclosure regulation: A review and suggestions for future research.” Working paper: University of Chicago.

38

Mansi, S. A., W. F. Maxwell, and D. P. Miller. 2004. “Does audit quality and tenure matter to investors? Evidence from the bond market.” Journal of ing Research 42: 755-793. Melumad, N. and L. Thoman. 1990. “On auditors and the courts in an adverse selection setting.” Journal of ing Research 28: 77-120. Pagano, M., Panetta, F. and L. Zingales. 1998. “Why do companies go public? An empirical analysis.” Journal of Finance 53: 27-64. Palmrose, Z-V. 1986. “Audit fees and auditor size: Further evidence.” ing Research 24: 97-110.

Journal of

Palmrose, Z-V. 1987. “Who got sued?” Journal of ancy 183: 67-69. Peel, M. J. and R. Roberts. 2003. “Audit fee determinants and auditor s: Evidence from the micro-firm sub-market.” ing and Business Research 33: 207-233. Pittman, J. A. and S. Fortin. 2004. “Auditor choice and the cost of debt capital for newly public firms.” Journal of ing and Economics 37: 113-136. Santos, J. A. C. 2003. “Why firm access to the bond market differs over the business cycle: A theory and some evidence.” Working paper: Federal Reserve Bank of New York. Seetharaman, A., F. Gul, and S. Lynn. 2002. “Litigation risk and audit fees: Evidence from UK firms cross-listed on US market.” Journal of ing and Economics 33: 91-115. Smith, C. W. and J. B. Warner. 1979. “On financial contracts and optimal capital structure: An analysis of bond covenants.” Journal of Financial Economics 7: 117161. Sunder, S. 2003. “Rethinking the structure of ing and auditing.” Yale ICF Working paper No. 03-17. Yale School of Management. Sweeney, A. 1994. “Debt covenant violations and managers’ responses.” Journal of ing and Economics 17: 281-308. St. Pierre, K. and J. A. Anderson. 1984. “An analysis of the factors associated with lawsuits against public ants.” The ing Review 59: 242-263.

39

Titman, S. and B. Trueman. 1986. “Information quality and the valuation of new issues.” Journal of ing and Economics 8: 159-172. Watkins, A. L., W. Hillison, and S. Morecroft. 2004. “Audit quality: A synthesis of theory and empirical evidence.” Journal of ing Literature 23: 153-193. Watts, R. L. 1977. “Corporate financial statements: a product of the market and political processes.” Australian Journal of Management 53-75. Watts, R. and J. Zimmerman. 1983. “Agency problems, auditing, and the theory of the firm: Some evidence.” Journal of Law and Economics 26: 613-634. Welker, M. 1995. “Disclosure policy, information asymmetry and liquidity in equity markets.” Contemporary ing Research 11: 801-828. Yu, F.

2005. “ing transparency and the term structure of credit spreads.” Journal of Financial Economics 75: 53-84.

Zingales, L. 2009. “The future of securities regulation.” Journal of ing Research 37: 391-425.

40

Table 1 Industry composition of the sample companies. UK SIC code 0001–0999 1000–1499 1500–3999 4000–4499 4500–4999 5000–5499 5500–5999 6000–6499 7000–7499 7500–7999 8000–9999

Industry Agriculture, hunting, forestry, and fishing Mining Manufacturing Electricity, gas, and water supply Construction Wholesale and retail trade Hotels and restaurants Transport, storage, and communication Business services Public services and defense Other service activities Total

Companies 79 12 719 3 748 1,221 148 338 1,414 3 454 5,139

41

Table 2 Variable means during the final year of the mandatory audit regime (2003) and the first year of the voluntary audit regime (2004).

RATINGit AGEit INTCOVit Salesit (£000) LTSit Assetsit (£000) LTAit LIQUIDITYit LEVERAGEit

Mandatory audit regime (t = 2003) Observations Mean 5,139 64.51 5,139 17.99 5,139 0.48 5,139 2,124.93 5,139 7.36 5,139 1,172.69 5,139 6.82 5,139 1.48 5,139 0.70

Voluntary audit regime (t = 2004) Observations Mean 5,139 68.37 5,139 18.99 5,139 0.46 5,139 2,077.30 5,139 7.33 5,139 1,173.06 5,139 6.84 5,139 1.58 5,139 0.69

Differences in means t-stat. = 9.32*** t-stat. = 3.09*** t-stat. = -1.03 t-stat. = -1.52 t-stat. = -1.31 t-stat. = 0.02 t-stat. = 1.07 t-stat. = 2.73*** t-stat. = -0.52

***, **, * = statistically significant at the 1%, 5%, 10% levels (two-tailed). Variable definitions RATINGit = the credit score (from 1 to 100 where a higher score implies a better rating) for company i in year t. AGEit = the age of company i in year t. INTCOVit = interest expense divided by earnings before interest and taxation (the INTCOVit ratio is capped at 2.00 and we assign a value of 2.00 if earnings before interest and taxation is negative). LTSit = log of total sales (£000). LTAit = log of total assets (£000). LIQUIDITYit = quick ratio ((current assets – inventory)/current liabilities). LEVERAGEit = total liabilities divided by total assets. LAFit = log of audit fees (£000). BIG4it = one if the company is audited by a Big Four audit firm, zero otherwise.

42

Table 3 Descriptive statistics after partitioning the sample by the company’s decision on whether to be audited (VOL_AUDITit) and the prevailing audit regime (mandatory audits in 2003, voluntary audits in 2004). A: The final year of the mandatory audit regime (t = 2003)

BIG4it Audit feesit (£000) LAFit RATINGit AGEit INTCOVit LTSit LTAit LIQUIDITYit LEVERAGEit

Companies wanting to be audited (VOL_AUDITit = 1) Observations Mean 3,440 0.08 3,440 5.68 3,440 1.52 3,440 67.48 3,440 18.81 3,440 0.50 3,440 7.40 3,440 6.91 3,440 1.51 3,440 0.68

Companies not wanting to be audited (VOL_AUDITit = 0) Observations Mean 1,699 0.02 1,699 4.27 1,699 1.26 1,699 58.50 1,699 16.32 1,699 0.42 1,699 7.28 1,699 6.64 1,699 1.41 1,699 0.74

Differences in means t-stat. = -8.16*** t-stat. = -11.87*** t-stat. = -13.32*** t-stat. = -13.64*** t-stat. = -5.11*** t-stat. = -3.50*** t-stat. = -4.21*** t-stat. = -12.80*** t-stat. = -1.87* t-stat. = 3.63***

B: The first year of the voluntary audit regime (t = 2004)

RATINGit AGEit INTCOVit LTSit LTAit LIQUIDITYit

Voluntarily audited companies (VOL_AUDITit = 1) Observations Mean 3,440 73.38 3,440 19.81 3,440 0.48 3,440 7.39 3,440 6.93 3,440 1.62

Unaudited companies (VOL_AUDITit = 0) Observations Mean 1,699 58.25 1,699 17.32 1,699 0.43 1,699 7.23 1,699 6.64 1,699 1.51

Differences in means t-stat. = -28.49*** t-stat. = -5.12*** t-stat. = -2.25*** t-stat. = -5.47*** t-stat. = -13.37*** t-stat. = -1.78*

43

LEVERAGEit

3,440

0.67 Table 3 (cont.)

1,699

0.75

t-stat. = 4.46***

C: Companies not wanting to be audited (VOL_AUDITit = 0)

RATINGit – RATING.t AGEit – AGE.t INTCOVit – INTCOV.t LTSit – LTS.t LTAit – LTA.t LIQUIDITYit – LIQUIDITY.t LEVERAGEit – LEVERAGE.t

Final year of the mandatory audit regime (t = 2003) Observations Mean 1,699 -6.01 1,699 -1.67 1,699 -0.05 1,699 -0.08 1,699 -0.19 1,699 -0.07 1,699 0.04

First year of the voluntary audit regime (t = 2004) Observations Mean 1,699 -10.13 1,699 -1.67 1,699 -0.03 1,699 -0.11 1,699 -0.20 1,699 -0.07 1,699 0.05

Differences in means t-stat. = -6.49*** t-stat. = 0.00 t-stat. = 0.77 t-stat. = 0.90 t-stat. = 0.36 t-stat. = 0.05 t-stat. = -0.49

D: Companies wanting to be audited (VOL_AUDITit = 1)

LAFit BIG4it RATINGit – RATING.t AGEit – AGE.t INTCOVit – INTCOV.t LTSit – LTS.t LTAit – LTA.t LIQUIDITYit – LIQUIDITY.t LEVERAGEit – LEVERAGE.t

Final year of the mandatory audit regime (t = 2003) Observations Mean 3,440 1.52 3,440 0.08 3,440 2.97 3,440 0.82 3,440 0.03 3,440 0.04 3,440 0.09 3,440 0.03 3,440 -0.02

First year of the voluntary audit regime (t = 2004) Observations Mean 3,440 1.53 3,440 0.07 3,440 5.00 3,440 0.82 3,440 0.02 3,440 0.05 3,440 0.10 3,440 0.04 3,440 -0.02

Differences in means t-stat. = 1.17 t-stat. = -0.92 t-stat. = 4.02*** t-stat. = 0.00 t-stat. = -0.51 t-stat. = 0.56 t-stat. = 0.27 t-stat. = 0.04 t-stat. = -0.41

44

***, **, * = statistically significant at the 1%, 5%, 10% levels (two-tailed). Table 3 (cont.) Variable definitions VOL_AUDITit = one if company i chooses to be audited during the voluntary audit regime, and zero if company i chooses not to be audited during the voluntary audit regime. BIG4it = one if the company is audited by a Big Four audit firm, zero otherwise. BIG4.t = the mean value of BIG4it across all companies in year t. LAFit = log of audit fees (£000). LAF.t = the mean value of LAFit across all companies in year t. RATINGit = the credit rating score (from 1 to 100 where a higher score implies a better rating) for company i in year t. RATING.t = the mean value of RATINGit across all companies in year t. AGEit = the age of company i in year t. AGE.t = the mean value of AGEit across all companies in year t. INTCOVit = interest expense divided by earnings before interest and taxation (the INTCOVit ratio is capped at 2.00 and we assign a value of 2.00 if earnings before interest and taxation is negative). INTCOV.t = the mean value of INTCOVit across all companies in year t. LTSit = log of total sales (£000). LTS.t = the mean value of LTSit across all companies in year t. LTAit = log of total assets (£000). LTA.t = the mean value of LTAit across all companies in year t. LIQUIDITYit = quick ratio ((current assets – inventory)/current liabilities). LIQUIDITY.t = the mean value of LIQUIDITYit across all companies in year t. LEVERAGEit = total liabilities divided by total assets. LEVERAGE.t = the mean value of LEVERAGEit across all companies in year t.

45

Table 4 Auditor choice and audit fees during the final year of the mandatory audit regime (2003).

Dep. var. =

VOL_AUDITit AGEit INTCOVit LTSit LTAit LIQUIDITYit LEVERAGEit BIG4it Industry dummy variables? R2 / pseudo R2 Observations

Auditor choice model BIG4it Model 1 Coefft. z-stat. 1.10 5.94*** -0.00 -1.00 0.29 3.87*** 0.15 2.35** 0.77 7.62*** 0.07 2.82*** 0.52 5.59***

YES 10.5% 5,139

Audit fees model LAFit Model 2 Coefft. t-stat. 0.13 7.20*** 0.01 5.64*** 0.13 11.55*** 0.25 21.83*** 0.19 14.84*** 0.01 1.74* 0.03 1.45 0.20 5.35*** YES 23.5% 5,139

Industry dummy variables are included (see Table 1) but the coefficients and the intercept are not tabulated. ***, **, * = statistically significant at the 1%, 5%, 10% levels (two-tailed). The t-statistics and zstatistics are reported in parentheses with standard errors that are adjusted for heteroskedasticity. Variable definitions BIG4it = one if the company is audited by a Big Four audit firm, zero otherwise. LAFit = log of audit fees (£000). VOL_AUDITit = one if company i chooses to be audited during the voluntary audit regime, and zero if company i chooses not to be audited during the voluntary audit regime. AGEit = the age of company i in year t. INTCOVit = interest expense divided by earnings before interest and taxation (the INTCOVit ratio is capped at 2.00 and we assign a value of 2.00 if earnings before interest and taxation is negative). LTSit = log of total sales (£000). LTAit = log of total assets (£000). LIQUIDITYit = quick ratio ((current assets – inventory)/current liabilities). LEVERAGEit = total liabilities divided by total assets.

46

Table 5 Credit ratings during the final year of the mandatory audit regime (2003) and the first year of the voluntary audit regime (2004).

Dep. var. =

VOL_AUDIT it AGEit INTCOVit LTSit LTAit LIQUIDITYit LEVERAGEit Industry dummy variables? R2 / pseudo R2 Observations

Final year of the mandatory regime (t = 2003) RATINGit Model 1 Coefft. t-stat. 7.68 13.41*** 0.25 14.57*** -10.21 -26.49*** 0.48 1.46 2.43 5.85*** 1.52 8.28*** -6.73 -10.33***

First year of the voluntary regime (t = 2004) RATINGit Model 2 Coefft. t-stat. 12.94 29.55*** 0.29 21.04*** -8.11 -25.82*** 1.82 7.30*** 3.79 11.08*** 0.28 2.51** -3.16 -6.83***

YES 27.4% 5,139

YES 23.6% 5,139

Industry dummy variables are included (see Table 1) but the coefficients and the intercept are not tabulated. ***, **, * = statistically significant at the 1%, 5%, 10% levels (two-tailed). The t-statistics and zstatistics are reported in parentheses with standard errors that are adjusted for heteroskedasticity. Variable definitions RATINGit = the credit rating score (from 1 to 100 where a higher score implies a better rating) for company i in year t. VOL_AUDITit = one if company i chooses to be audited during the voluntary audit regime, and zero if company i chooses not to be audited during the voluntary audit regime. AGEit = the age of company i in year t. INTCOVit = interest expense divided by earnings before interest and taxation (the INTCOVit ratio is capped at 2.00 and we assign a value of 2.00 if earnings before interest and taxation is negative). LTSit = log of total sales (£000). LTAit = log of total assets (£000). LIQUIDITYit = quick ratio ((current assets – inventory)/current liabilities). LEVERAGEit = total liabilities divided by total assets.

47

Table 6 Changes in credit ratings when auditing switches from being mandatory to voluntary (2003–2004). Full sample Dep. var. = Δ(RATINGit – RATING.t) Model 1 Coefft. t-stat. Intercept -3.91 -8.76*** VOL_AUDITit 5.84 10.61*** -8.16 -23.76*** Δ(INTCOVit – INTCOV.t) -0.04 -0.06 Δ(LTSit – LTS.t) 2.84 2.91*** Δ(LTAit – LTA.t) -0.11 -0.48 Δ(LIQUIDITYit – LIQUIDITY.t) -2.58 -2.84*** Δ(LEVERAGEit – LEVERAGE.t) R2 / pseudo R2 Observations

13.8% 5,139

Companies that switch from mandatory audits to voluntary audits Δ(RATINGit – RATING.t) Model 2 Coefft. t-stat. 1.93 5.99*** -7.90 0.34 2.24 -0.01 -1.99

-19.07*** 0.41 1.81* -0.04 -1.50 11.5% 3,440

Companies that switch from mandatory audits to no audits Δ(RATINGit – RATING.t) Model 3 Coefft. t-stat. -3.91 -8.81*** -8.81 -1.01 3.90 -0.36 -3.34

-14.33*** -1.01 2.35** -0.68 -2.93***

13.4% 1,699

***, **, * = statistically significant at the 1%, 5%, 10% levels (two-tailed). The t-statistics and z-statistics are reported in parentheses with standard errors that are adjusted for heteroskedasticity. Variable definitions: VOL_AUDITit = one if company i chooses to be audited during the voluntary audit regime, and zero if company i chooses not to be audited during the voluntary audit regime. RATINGit = the credit rating score (from 1 to 100 where a higher score implies a better rating) for company i in year t. RATING.t = the mean value of RATINGit across all companies in year t. AGEit = the age of company i in year t. AGE.t = the mean value of AGEit across all companies in year t. INTCOVit = interest expense divided by earnings before interest and taxation (the INTCOVit ratio is capped at 2.00 and we assign a value of 2.00 if earnings before interest and taxation is negative). INTCOV.t = the mean value of INTCOVit across all companies in year t. LTSit = log of total sales (£000). LTS.t = the mean value of LTSit across all companies in year t. LTAit = log of total assets (£000). LTA.t = the mean value of LTAit across all companies in year t. LIQUIDITYit = quick ratio ((current assets – inventory)/current liabilities). LIQUIDITY.t = the mean value of LIQUIDITYit across all companies in year t. LEVERAGEit = total liabilities

48

View publication stats

divided by total assets. LEVERAGE.t = the mean value of LEVERAGEit across all companies in year t.

49