Are Modified Audit Opinions Related to the Availability of Credit? Evidence from Finnish SMEs

Abstract

We study the association between credit availability and modified audit opinions using a sample of more than 50,000 observations for small- and medium-sized companies. Studies in finance suggest that companies use trade-credit as a source of financing when institutional debt is not available (e.g. Petersen and Rajan, 1994; Danielsson and Scott, 2004). Building on these studies, we study whether modified audit opinions are associated with an increased use of trade credit relative to bank debt. We find no association between modified audit opinions and our measure of credit rationing. Our archival evidence focusing on SMEs is contrary to much of the earlier research finding that modified audit opinions provide incremental information for lenders. Our study adds to the scarce literature on the role of audit reports as a source of information in SME finance.

Acknowledgement:

We wish to thank Tomi Seppälä, Aasmund Eilifsen, Juha Kinnunen, Ann Vanstraelen (associate editor), two anonymous reviewers and participants at the 2007 European Accounting Association Meeting for helpful comments and suggestions. Stefan Sundgren gratefully acknowledges financial support from the Academy of Finland (grant 126630). Lasse Niemi is thankful for the financial support from the Foundation of the Helsinki School of Economics.

Notes

Paper accepted by Ann Vanstraelen.

Currently, the global private sector output and employment generated by SMEs is estimated at 52% and 67%, respectively (ACCA, 2010:31). Due to the significant contribution of SMEs to the economy, politicians and leaders in Europe and elsewhere have made attempts to facilitate better access to finance for SMEs. A recent example of these efforts is a permanent SME Finance Forum which was set up by The European Commission to encourage new approaches to improve access to finance for SMEs (SME Finance Forum, 2010).

A red flag for liquidation is a type of going-concern opinion that is issued if losses have consumed shareholders' total equity beyond a certain point.

Studies of the self-fulfilling prophecy (e.g., Gaeremynck and Willekens, 2003) suggest that going concern modified audit opinions cause bankruptcy and a possible reason for this is that providers of debt capital react adversely to modified audit opinions, which indicate going concern problems.

According to the current Auditing Act, a statutory audit is required if two out of three size thresholds are exceeded: net sales 0.2 million, total assets 0.1 million, and the average number of employees 3. In other words, size thresholds in the EU directive for net sales and total assets are 44 times higher than those in the current Auditing Act. The Act came into force in 2007.

Furthermore, at least in the short run companies can obtain temporary financing by simply not paying trade credits. Trade creditors generally have low priority in bankruptcy, and thus, a bankruptcy petition is not a credible threat if a company delays payment. Using a sample with 761 small and middle-sized Finnish companies, Bergström et al. (2004) found that trade creditors and other general creditors on average receive only 4% of their claims (median 0%).

Before joining the EU in 1995, Finland harmonised its laws regarding financial reporting and auditing with EU legislation through the new Accounting Act and Auditing Act. After that, changes in EU legislation have been implemented relatively quickly in Finnish legislation.

According to the report (European Commission, 2003: 29-30), the payment periods differ by country. In Italy, for example, it takes on average 87 days before payment is made, while Swedish firms collect their debts within an average of 34 days. In our setting (Finland) the typical contractual payment period is between 14 and 30 days, which is closer to that in Sweden than Italy.

Finland, Denmark, and Sweden have exempted the smallest companies from the audit requirement. In Finland this change took place in 2007 (Auditing Act 13.4.2007/459).

Subsection 19 of the Auditing Act (936/94, Chapter 4) stipulates the contents of the audit report. According to subsection 19, the auditor shall provide an opinion on whether financial statements are prepared according to the Accounting Act and other prevailing laws and instructions, and whether financial statements give a true and fair view. The audit report also has to contain an opinion regarding the proposal for profit distribution made by the board of directors. Subsection 19 explicitly states that if there have been any acts or negligence by the board members or the managing director that could give rise to claims for damages, or if the board members or the managing director have violated any applicable laws or company by-laws, a qualified audit opinion has to be rendered.

In June 2007 the new Auditing Act (13.4.2007/459) replaced the former Auditing Act (936/94) that prevailed for the years in our analyses. Under the new law the auditor reporting focuses more clearly on financial statements. The requirements of opinions on the board's liability and proposition on the profit distribution in a standard report have been removed from the new Auditing Act. The new Auditing Act, however, has not changed non-standard audit reports. The auditor shall still provide a qualified audit report, if the board has not followed the rules of law or by-law (cf. ISA 250). Also, the Company Law has changed recently. For the relevant changes from the paper's point of view, see footnote 12.

The Pearson correlation matrix of the alternative credit availability measures is presented in Appendix 1. An examination of the matrix shows that correlations are high between measures aimed to measure the same underlying construct. For example, the correlation between UAP and ΔTCREDIT/ASSETS is 0.81 and both these variables are intended to measure an unexpected increase in the use of trade credit. The correlations between INSTDEBT and the trade credit based measures are low. The exact calculations of the measures are presented in Table 1.

As can be seen in Table 1, we calculate the expected change in trade credit as the amount at the year with the modified audit opinion multiplied with the increase in revenues. Marquardt and Wiedman (2004) assume that the expected change is a function of the change in cost of goods sold. The cost of goods sold is not available for the firms in the data. However, the fact that gross margins are quite constant in companies over the years implies that the use of revenues should give approximately similar results to those obtained from the use of cost of goods sold.

The new Company Law (21.7.2006/624) (CL) came into force in September 2006. Although the personal obligation of board members to initiate liquidation themselves has been removed, the requirement to inform the owners when the total equity is less than half of the shareholder equity is still in the new CL. Under the new law the public register (PRH) must also be informed if the losses have consumed the entire equity (negative total equity). Like any violations of the CL, the auditor has to report if the company has not informed the register of the negative equity. Now it is more clearly the responsibility of the owners to start the liquidation process if the lack of equity is not remedied in due time. One plausible reason underlying recent changes in the CL and other relevant laws and rules is the recent development at EU level to reduce the administrative burden and costs for SMEs. Liquidation of a company is a time-consuming and laborious legal undertaking, and therefore may be very costly for those responsible for it.

The requirement of retained earnings has been removed from the new CL. This change relates to the changes in the restrictions on dividend payments. Earlier, dividends could be paid only if there were sufficient retained earnings (or some special reserve funds), whereas now the restrictions concentrate on the company's liquidity instead of the amount of equity. The current CL explicitly states that dividend payments or related-party loans shall not jeopardize the company's ability to pay its debts. Also, related party loans exceeding 20,000 euros must be reported in the financial statements (section 8, sub section 6.1). Moreover, if the loan to related parties violates any of the requirements in the CL, these violations must be described in the footnotes to the financial statements. As the reporting requirements have been increased in the new CL and the auditor has to see that financial statements provide a true and fair view, the changes regarding auditor reporting are somewhat cosmetic in practice.

Correspondingly, companies with income statements classified by function were excluded from the sample. However, Finnish SMEs seldom have income statements in which costs are classified by function.

Indeed, companies may be without institutional debt either because they do not need any debt or because institutional lenders do not grant them credit. If modified audit opinions have an adverse impact on credit availability, one would expect that a change from having institutional debt to not having institutional debt is more common for companies with a modified audit opinion than for companies with a clean audit opinion. In order to provide some evidence of this, we calculated the percentages of companies that switched from having institutional debt to not having institutional debt and studied whether this was more common among companies with modified audit opinions. Information about audit opinions for year t-1 and institutional debt for year t-1 and year t is available for 170,915 firm-years. Of these firm-years, 10,297 (6.0%) switched from having institutional debt to not having institutional debt and 7,950 (4.6%) switched from not having institutional debt to having institutional debt. The results showed that 4.15% of the companies that switched from having institutional debt to not having institutional debt had a modified audit opinion, that 4.00% of the companies that remained in the same category had a modified audit opinion and that 3.70% of the companies that switched from not having institutional debt to having institutional debt had a modified audit opinion. A Pearson Chi-square test showed that there is no significant association between a modified audit opinion and a switch from having institutional debt to not having institutional debt (p-value 0.451).

627 firm-years with modified audit opinions and 24,443 firm-years with clean audit opinions remained as we required that the audit report should be clean during the two years before inclusion in the sample. We ran regressions with the same variables as in Table 4 and RED FLAG and FIN STAT had insignificant coefficients in the regressions. LENDING RULES had a significant negative coefficient in these regressions, not the predicted positive one.

Even if we think it is more appropriate to test our prediction on a sample with firms having institutional debt, we also run Models 1 to 4 on all firms for which the necessary variables are available. The sample included 88,821 observations for regression one and, as in Table 4, somewhat fewer for regressions two to four. RED FLAG had positive coefficients significant at the 0.10 level in regression one and at the 0.05 level in regression four. However, the magnitudes of the variables are very small even if they are significant. For example, the coefficient in regression four is only 0.009. To conclude, with a few exceptions all modified audit opinion variables are insignificant in the regressions.

We considered a continuous variable calculated as debt at year t + 1 less debt at year t, divided by total assets as a supplementary measure (ΔDEBT/ASSETS). We discuss results with INSTDEBT in text because ΔDEBT/ASSETS is influenced by observations for which debt is decreasing and this decrease is a consequence of the maturity of previously granted loans and not the availability of credit after the modified audit report. Thus, INSTDEBT should be a better measure of credit availability than ΔDEBT/ASSETS. However, the results on all sub-samples were qualitatively similar with ΔDEBT/ASSETS as with INSTDEBT.

Marquardt and Wiedman (2004) relate unexpected changes in trade-credit to earnings management. However, as can be seen from Table 2, companies with the corresponding types of modified audit opinions have a higher leverage and lower current ratio than the companies with a clean audit opinion. Furthermore, the firm-years with a modified audit report indicating non-compliance with Finnish GAAP have a lower NITA than the firm-years with clean audit opinions. Hence, it is unlikely that the companies in our sample on average have the incentive to manage earnings downward and that earnings management by the companies with a modified report would be the reason for the association between UAP and modified audit reports.

A firm was included in the sample if the audit report at t-1 was clean and modified at year t. Each firm was included only once in the sample. Furthermore, for inclusion in the sample we required that the credit availability measures and all explanatory variables in the empirical model except GROWTH should be available. The reason why the availability of GROWTH was not required was that it would have reduced the sample size. Furthermore, we indirectly control for the growth in the regressions via the inclusion of LNASSETS at years t-1 and t in the regressions reported in Panel B of Table 5.

The rationale behind this comparison is as follows. Days payable may change for natural reasons as well as a consequence of credit rationing from institutional lenders. Companies that have institutional debt typically have to annually repay a fraction of their loans. Consequently, if a modified audit opinion is associated with an increased likelihood of credit rationing from institutional lenders and trade credit is used as a last source of financing, one would expect a greater increase in the usage of trade credit after a modified audit opinion for companies with institutional debt than for companies without institutional debt.