«John Christian Langli and Tobias Svanström © John Christian Langli and Tobias Svanström 2013. All rights reserved. Short sections of text, not to ...»
The results in Gaeremynck and Willekens (2003) are from a period when the going concern standard was principle-based, allowing auditors considerable discretion. The standard was replaced by a rulebased standard in 2000. Carello, Vanstraelen,and Willenborg (2009) analyze 132 firms filing for bankruptcy during 1995-1996 and 89 firms filing for bankruptcy during 2001-2002, and match these firms with an equal number of financial stressed non-bankrupt firms using year, industry and probability of bankruptcy. Carello et al. (2009) find that non-Big N auditors committed more type II errors than Big N auditors when the standard was principle-based, but not after introduction of the rule-based standard. Furthermore, after introduction of the rule-based standard, they document an increase in type I errors, a decrease in type II errors and that there is no longer a negative relationship between going concern reports and an auditee’s size (which is found in almost all previous studies). In conclusion, they state that “the Belgian standard tends to favor creditors and disfavor auditors, and perhaps also companies and employees. Evaluation of the net of these effects depends on the priorities one assigns to the parties, an evaluation that we respectfully leave to Belgian standard-setters and regulators.” (p. 1425).
Regulatory reviews of auditors The positive association between audit quality and the size of the audit firm/office/team that have been documented for public firms (Francis and Yu 2009; Choi et al. 2010) might be even more apparent in audits of private companies since the audit is conducted, at one extreme, by auditors employed by Big N firms and, at the other extreme,by sole practitioners working alone. In these very different work environments, the levels of internal quality control systems, external input from colleagues and the auditors’ incentives vary significantly. Sundgren and Svanström (2012a) is the only study we are aware of that analyzes regulatory reviews of auditors in private companies. The researchers find that auditors working at offices with three or more CPAs were significantly less likely to be subject to disciplinary sanctions compared with auditors at offices with only one or two CPAs.
The significant positive association between audit office size and audit quality was found for non-Top 6 audit firms only (top 6 = Big 4 +BDO + Grant Thornton), suggesting that audit quality is more heterogeneous among small audit firms than among large audit firms and that small offices more often fail to meet the quality requirements.
The Big N fee premium There is ample evidence using public firm data that large auditors charge significantly higher fees than smaller auditors. The fee premium that large audit firm earn can be caused by several (nonindependent) reasons, such as lack of competition among auditors, which is a primary concern among regulatory bodies; reputation and brand name effects; higher quality audits; higher production cost, better trained staff and more advance technology; and higher potential losses in the event of litigation (large audit firms have “deep pocket”) (Chaney et al. 2004, Clatworthy et al. 2009).
The existence of a fee premium has also been studied using samples of private firms, which has the advantage of reducing the effects that “deep pockets”, litigation risk, loss of reputation, and market concentration may have on fees. Thus, if the fee premium exists, the likely causes are not related to these effects.
Chaney et al. (2004) use a large sample of UK private firms and find a fee premium for Big N firms using standard OLS-regressions, but no fee premium when they control for potential self-selection
using the Heckman approach:
“[The] results are consistent with the notion that auditees, when not compelled by market pressures to choose a Big 5 auditor, choose the lowest-cost auditor available; further, our results suggest that clients in our setting, on average, do not view Big 5 auditors as superior in terms of the perceived quality of services provided to a degree sufficient to justify a fee premium.” (p 70).
Lennox, Francis and Wang (2012) raise doubt about the results in Chaney et al. (2004) due to how the Heckman test is implemented. Clatworthy et al. (2009), also using a large UK private firm sample but controlling for self selection bias using different methods, find results that directly contradict the findings in Chaney et al. (2004). They document a fee premium for Big 4 firms and find no evidence supporting the view that Big 4 clients choose the lowest cost auditor. Price premia for Big N auditors are also found in Belgium (Willekens and Achmadi 2003) and Norway (Hope and Langli 2010, Hope et al. 2012), and for the Top-6 auditors in Sweden (Sundgren and Svanström 2012a).
Access to credit Most private companies need external financing, and their ability to raise equity may be limited by e.g. insufficient private wealth that the owners can invest in the firm or reluctance to open up for new owners. Therefore, gaining access to credit, primarily through banks, is often of vital importance.
Audited accounting information and personal interviews are two important sources of information used by lenders (Berry and Robertson 2006). In the presence of asymmetric information, lenders may respond to uncertainty about a borrower’s creditworthiness by simply not accepting loan applications. Uncertainty may be reduced for firms that engage an auditor in a voluntarily audit regime, voluntarily hire a high quality auditor or two auditors in regimes where one auditor is mandatory, but may increase for firms that receive modified audit opinions.
Hope, Thomas, and Vyas (2011: 937) use survey data from the World Bank (close to 50,000 manufacturing and service firms from 68 countries) and find that audited financial information “is associated with lower perceived financing constraints”. Allee and Yohn (2009), using private US firm data (see below), find that firms with audited financial statements have lower probabilities of getting a loan denial compared to firms with non-audited financial statements. Thus, both studies imply that auditing eases firms’ access to credit. Niemi and Sundgren (2012) analyze if modified audit opinions reduce the likelihood of obtaining credit from institutional lenders among Finish SMEs and thus increase the use of trade credit, but they fail to find any relationship. The non-importance of modified audit opinion supports the experimental evidence in Wright and Davidson (2001) who analyze the effect of auditor attestation on commercial lending decisions. They found that bankers’ risk assessment and loan approval decisions for a privately-held company in the wholesale cleaning supply industry were not affected by whether there was no attestation, a review or an audit.
Cost of debt One effect of reduced agency costs from auditing is that it reduces financiers’ lending costs. In a competitive market, this reduction will be passed on to lenders. Thus firms that voluntarily engage an external auditor are expected to obtain loans at lower interest rates than firms without an auditor.
Blackwell et al. (1998) analyzed the association between actual interest rates paid and the degree of external verification of the borrowers’ financial statement for 212 small, private US firms, and found that audited companies pay lower interest rates than unaudited companies after controlling for firmspecific risk factors and relevant loan characteristics. The size-matched sample showed that the interest rate of audited companies was 25 basis points lower than those of unaudited companies.
The estimated interest saving was 30 to 50 per cent of audit fees paid. Minnis (2011) also finds significantly lower interest rates for audited firms using a large sample of private US firms (25,784 firm-year observations from 12,616 unique firms from 2001-2007): The interest rate is on average 69 basis point lower for audited firms (varying between 25 and 105 basis point depending on model), which amounts to an interest saving of 25,000 USD (or 6 per cent of net profit before tax) for the average sample firm. However, the US evidence is mixed, as Allee and Yohn (2009) and Cassar (2011), using data collected by the Federal Reserve Board in 2003 (the National Survey of Small Business Finances), do not find lower interest rates for firms with audited financial statements. These researchers use the interest rates on the firms’ most recent loans and they control for, among other things, the terms of the loan (e.g. fixed or floating rate and whether there is collateral for the loan).
Lack of support for better terms on debt is also found by Fortin and Pittman (2007) who analyze the yield spreads and credit ratings on bonds issued by private firms. This is contrary to findings for public firms, which are that Big N auditors are associated with lower cost of debt (see De Franco et al.
2011) Kim et al. (2011) exploit the Korean environment, where all firms except those with total assets less than approximately USD 7 million are required to have their financial statements audited, to assess the effect the cost of debt. Using a large sample (72,577 firm-year observations from the years 1987they find a larger reduction in interest rates than those reported by Blackwell et al (1998).
Depending on the estimation method used, the average interest cost savings from a voluntary audit range from about 56 to 124 basis points. They also report a significant reduction in interest rates for those that engage an auditor for the first time.
The studies by Blackwell et al (1998), Kim et al. (2011) and Minnis (2011) analyze the effect on interest rates in environments where auditing is voluntarily. Karjalainen (2011) analyzed a sample of 3,890 unique Finnish firms (10,799 firm-year observations from 1999-2006) that all were subject to statutory auditing. Subject to firm size, Finish firms may choose between auditors with no professional certification, HTM-auditors (which are regarded as the second-tier auditors) and KHT auditors (which are regarded as first-tier auditors). The Big N firms are also present in Finland and firms may also choose to have multiple auditors. Thus, there are different sets of auditors to choose from that have different degrees of perceived quality. Karjalainen (2011) documents that the presence of Big N-auditors or multiple auditors is associated with lower cost of debt. Karjalainen (2011) also finds that firms receiving modified audit opinions have higher interest rates (and lower accruals quality) than firms with clean opinions. Using Spanish data, Cano-Rodriguez and Alegria (2012) also find that Big N auditors are associated with lower interest rates for private firms (but not for public firms).
Credit ratings Kim et al. (2011) and Minnis (2011) among others compute firms’ interest rates using information from the income statement and the balance sheet. Inferring interest rates from financial statements has several potential problems, see Cassar (2011) for a thorough discussion. To illustrate, it is more likely that unaudited firms have unrecognized liabilities compared to audited firms, since auditors perform checks that all liabilities are recognized with the correct amounts while no such controls are performed in unaudited statements. Thus, since liabilities are the denominator in the interest rate calculation, audited firms may by construction get lower interest rates and this may explain the reduction in interest rates that are documented in studies using interest rates inferred from financial statements.
Lennox and Pittman (2011) circumvent the problems with interest rate calculations by using credit rate scores as the dependent variable. Their purpose is to analyze if statutory auditing suppresses valuable information because a firm’s ability to signal low credit risk is reduced. In statutory audit regimes, firms may choose a high quality auditor or not, while in voluntary audit regimes they also have the option to not hire an auditor. Lennox and Pittman (2011) utilize the change in audit regulation in the UK that made it possible for a number of private firms to opt out of an audit for the first time in 2004 (i.e. auditing was statutory for all sample firms in 2003, but not in 2004). The results show that firms that remain audited get a significant upgrade in credit rating while those that opt out get an even larger downgrade. Since there is no change in the assurance effect of auditing for those that remain audited, the upgrade can be attributed to the signaling effect for firms that for the first time are able to signal their willingness to be audited. The opt-out companies were less likely to appoint Big N auditors and paid lower audit fees under the mandatory regime relative to the companies that remained audited. The authors suggests that the benefits from requiring these companies to be audited are likely to be small since they privately contract for a low level of audit assurance when audits are legally required.
The notion that auditing matters and is valued by credit rating agencies is further supported by Zerni, Haapamäki, Järvinen and Niemi (2012) who find among other things that private Swedish companies that engage two auditors have a better credit rating than companies hiring ‘only’ one auditor.