The starting point is bankruptcy data for the period 2012–2020 from the Reach database of Bureau van Dijk and Company.info, encompassing all bankruptcy-related deregistrations during this period. To ensure comprehensive coverage of organizations subject to statutory audit, I cross-referenced these bankruptcy records with data from Company.info, which provides information on auditor involvement and filing obligations for Dutch legal entities. This initial dataset comprised 2,043 unique bankruptcies.

To refine the sample and focus exclusively on organizations subject to mandatory statutory audit, I implemented several filtering criteria. First, I excluded small legal entities and specific legal forms not subject to mandatory audit under Dutch law on the supervision of audit firms (Wet toezicht accountantsorganisaties). Second, I focused on organizations required to file income statements in the Netherlands, as these are generally subject to mandatory audit requirements. Third, I applied size criteria from the Dutch commercial code, filing size reported to the Chamber of Commerce, and verified the involvement of auditors operating under the statutory audit framework of the AFM (Dutch Authority for the Financial Markets) audit oversight. This process resulted in a final sample of 572 bankruptcies between 2012 and 2020, representing organizations that were subject to a mandatory audit in at least one of the three years preceding bankruptcy (Table 1). Most insolvencies in the sample were in the wholesale and retail, manufacturing, and construction industries, respectively (Table 2). While this selection procedure aims for precision, some organizations marginally outside the statutory audit scope may have been included, potentially leading to a conservative overestimation of bankruptcies strictly subject to mandatory audit. I acknowledge this as a potential limitation.

For each of the 572 bankruptcies, I collected financial statements and audit opinions for the three fiscal years preceding bankruptcy. The primary data sources were Company.info, the Reach database from Bureau van Dijk, and the Dutch trade register. I prioritized obtaining the originally filed PDF documents to ensure data accuracy. In cases where financial statements were unavailable from these sources (11 bankruptcies), those cases were excluded from the final sample to maintain data completeness.

Financial statement data were extracted from the Reach database and supplemented with hand-collected data directly from the original filed financial statements to ensure accuracy and capture nuanced information, particularly regarding management disclosures on going concern. Information on audit opinions was collected using an automated data-gathering algorithm to enhance efficiency. To validate the algorithm’s performance and ensure data reliability, I independently hand-collected audit opinion data and reconciled all discrepancies. Furthermore, extensive logical checks were performed on the hand-collected data, and a colleague re-performed a fifth of the hand-collected work as a quality review. Insolvency claims data, including claims from tax authorities and the employee insurance agency (‘UWV’), were hand-collected from the most recent insolvency reports filed at the courts, obtained from Faillisementsverslagen.com and Company.info as of summer 2021.

To study the reporting behaviour of Dutch organizations under statutory audit requirements in the three years preceding bankruptcy, my analysis focuses on several key variables:

I employed descriptive statistics to analyse filing behaviour, management disclosures, and audit opinion types in the three years preceding bankruptcy. I calculated frequencies, percentages, means, and medians to summarize the data. To estimate GCO error rates, I combined the bankruptcy data with aggregate market data on statutory audits and GCO issuance rates obtained from the AFM Monitor. This allowed me to contextualize the findings within the broader Dutch audit market and provide estimates of Type I and Type II misclassification rates at the market level. For the analysis of societal impact, I calculated aggregate and average figures for jobs lost and insolvency claims. I provided descriptive statistics on the top 25 largest bankruptcies by claim size to investigate the skewed distribution of bankruptcy damages.

The analysis of filing behaviour in the three years before bankruptcy reveals alarmingly low rates of timely and complete financial statement filings, indicating systemic issues of regulatory compliance among companies nearing insolvency. Figure 1 and Table 3 illustrate this. In the critical year immediately preceding bankruptcy, just 12 percent of organizations filed timely audited financial statements or a valid exemption. While compliance rates were higher in the years further removed from bankruptcy (56 percent and 64 percent for years two and three, respectively), the drop-off in the year before failure is concerning.

The significantly lower filing rate in the immediate pre-bankruptcy year (84% non-filing) compared to earlier years (10% and 4% non-filing in years two and three) suggests that the impending bankruptcy likely disrupts the financial reporting and audit process. However, this disruption is symptomatic of a deeper problem: a lack of proactive measures to address regulatory compliance even as financial distress becomes imminent. Furthermore, the composition of non-compliance reveals that it is not simply a matter of late filings but also involves the filing of unaudited draft financial statements and, critically, the absence of audit opinions in many cases, eroding the reliability and credibility of the limited filings that do occur.

Figure 1 visually emphasizes the decline in filing compliance as bankruptcy approaches, underscoring the need for regulatory intervention to improve filing discipline, particularly for companies in financial distress. As Table 3 details, a significant portion of non-compliance stems from filing past the twelve-month deadline and the filing of draft financial statements (voorlopige jaarreke­ning), neither of which meets the statutory requirements for transparency and accountability.

Moreover, for the subset of companies that do file, the reporting lag is excessively long, averaging 295 days (median 303 days). This delay further diminishes the timeliness and relevance of financial information for stakeholders attempting to assess and respond to emerging risks. The delayed filing of financial statements, or their complete absence, should be considered a critical “red flag” signalling potential financial distress and a lack of accountability to stakeholders.

An essential aspect of corporate accountability is management’s responsibility to provide transparent and informative disclosures about the company’s financial condition, particularly regarding going concern risks. However, my findings, depicted in Figure 2, reveal a significant deficiency in management’s disclosure of discontinuity risks in pre-bankruptcy filings. Only 29 percent of adopted financial statements for the year before bankruptcy include any explicit disclosure of continuity risks beyond a standard going concern assumption statement. This limited disclosure represents a failure by management to inform stakeholders of potential financial difficulties proactively and undermines the principle of transparent and accountable financial reporting. It is important to note that companies may face strong disincentives to disclose discontinuity risks, such as fear of damaging relationships with key stakeholders, including suppliers, customers, or lenders.¹

The majority of adopted financial statements (71 percent) remain silent on continuity risks, even as bankruptcy looms. This lack of disclosure raises serious questions about management’s adherence to accounting standards and their commitment to providing stakeholders with a fair and accurate view of the company’s financial prospects. In light of this finding, the Brydon (2019) report’s call for a mandatory “resilience statement” with explicit going concern assessments by management gains further credence.

The audit opinion filing rate in the year before bankruptcy is even more concerning than the overall financial statement filing rate. Figure 3 demonstrates a dramatic drop in audit opinion filings to a mere 11 percent in the year before bankruptcy, compared with 65 percent and 79 percent in the two years prior, respectively. This near-absence of audit opinions in the critical pre-bankruptcy year severely limits the credibility of financial information available to stakeholders when they need it most. It suggests that the audit process was not completed for companies approaching insolvency. Since no public data is filed, it is hard to attribute this to either the auditor or the client. However, some insolvency collectors publicly report on this process after the bankruptcy. For example, the insolvency collector of Mexx reports that its auditor, Deloitte, privately warned Mexx’s management on June 30, 2014, regarding the going-concern risk, eight months before Mexx filed for bankruptcy on March 13, 2015. Yet, in the trade register, only the unaudited financial statements for 2013 were filed on September 5, 2014.

Other companies file draft financial statements stating that the auditor is still considering the going-concern risk. For example, Partner Logistics Europe B.V. states in its draft 2010 financial statements, filed on January 27, 2012, before its bankruptcy in June 2012: “The auditor is considering his position on the paragraph regarding the continuity of the company that needs to be included in this report.” Another example is Rolsa N.V. which states: “Due to the lack of clarity with regards to the going concern of the Company, the Company has not been able to finalise the audited accounts.” In these three examples, the auditor correctly identified the material going-concern uncertainty, yet it was not always transparent to stakeholders. In the cases of Partner Logistics and Rolsa, the financial statement users at least had a disclosure by the company. Still, in the case of Mexx or when no draft financial statements are filed, public information is lacking.

Since 2023, the auditor has the obligation, for statutory audits, to promptly inform the workers’ council when a material uncertainty relating to going concern is included in the audit opinion. However, this applies only when the audit opinion has already been issued. Instances where the auditor does not issue an opinion because the situation is too uncertain or unclear remain undisclosed to the public.

Furthermore, when audit opinions are filed in the year before bankruptcy, the majority fail to address going-concern uncertainties adequately. Of the audit opinions filed for the pre-bankruptcy year, 63 percent (39 out of 62) did not include a Going Concern Opinion (GCO), constituting Type II misclassifications (failures to warn). This high rate of Type II misclassifications, although based on a small number of filed opinions, points to a potential weakness in auditor going concern assessments in the Netherlands. It raises concerns about the effectiveness of audits in providing timely and reliable signals of impending financial distress. These results are consistent with prior anecdotal evidence by Litjens (2010).

Besides whether the auditor issued a GCO, I am also interested in modified audit opinions (MAOs). MAOs include qualified audit opinions, disclaimer of opinions, and adverse audit opinions. Figure 4 presents data on Modified Audit Opinions (MAOs) beyond GCOs. While MAOs are more frequent than GCOs, their prevalence also declines in the year before bankruptcy. However, the relatively higher incidence of MAOs compared to GCOs in all periods suggests that auditors are identifying issues with financial statements and internal controls, but are perhaps hesitant to issue GCOs, potentially due to the concerns about Type I misclassifications and self-fulfilling prophecies, even though my analysis casts doubt on the magnitude of this self-fulfilling prophecy risk at a macro level.

Besides a GCO or MAO the auditor can also include a key audit matter (KAM) in its audit opinion. The auditor might inform stakeholders in a KAM about the discontinuity risk and how the audit firm addressed it. KAMs are voluntary for non-public interest entity (PIE) audits, and I identified 14 KAMs in the audit opinions in the three years before bankruptcy. From those KAM’s only half discussed the discontinuity risk. The limited number of Key Audit Matters (KAMs) addressing discontinuity risk, even when MAOs are issued, further suggests a lack of transparent communication about going concern issues.

The significant societal impact of bankruptcies underscores the importance of effective going concern reporting. Table 4 provides a detailed overview of the top 25 bankruptcies by insolvency claims in the Netherlands. These 25 bankruptcies alone account for EUR 11.4 billion in insolvency claims and involved over 43,000 employees. The largest bankruptcy, Petroplus International B.V., involved claims exceeding EUR 3.7 billion. These figures highlight the immense economic disruption associated with a relatively small number of large bankruptcies.

My broader analysis, summarized in Figure 5, estimates that bankruptcies in the sample affected 127,670 jobs and EUR 15.7 billion in insolvency claims. These figures, which include substantial amounts owed to tax authorities and the employee insurance agency, further emphasize the significant societal burden imposed by corporate failures. The skewed distribution of bankruptcy damages, with a few large insolvencies accounting for the majority of losses, highlights the need for effective early warning mechanisms, such as timely and reliable going concern reporting, to mitigate these large shocks.

Table 5, Figure 6, and Figure 7 present the analysis of GCO misclassification or error rates. I identify a relatively low Type II error rate of 0.03 percent, based on 39 identified events. Each Type II misclassification represents a potentially avoidable instance of unexpected bankruptcy, with its associated stakeholder losses and societal costs. The fact that the last Type II error I identified was for an audit opinion issued in 2019 suggests that this is not merely a historical issue, but a persistent challenge. However, it is important to note that a Type II misclassification is not necessarily an “error”, as new information might come to light after issuance of the audit opinion, that casts doubt on the ability of the company to continue as a going concern, that the auditor could not have known when issuing the opinion, even when performing a perfect audit.

Conversely, I estimate a high Type I error rate of 99.26 percent for the 12 months after balance sheet date. However, a longitudinal analysis examining bankruptcy proceedings within 36 months of GCO issuance reveals that, in 92.16 percent of cases, bankruptcy proceedings do not follow a GCO within this timeframe. This evidence, coupled with the increasing trend in GCO issuance over time (Figure 7), suggests that concerns about the size of the self-fulfilling prophecy effect of GCOs, at a macro level, are likely overstated. Auditor conservatism, while potentially contributing to Type I misclassifications, does not appear to be a primary driver of macro-level bankruptcies in the Netherlands. Instead, the more pressing concerns are the relatively low rate of GCO issuance in the year before bankruptcy and the potential for Type II misclassifications, which raise issues of auditor accountability and missed opportunities for early warning to stakeholders.

It is important to acknowledge the limitations of this research, which also pave the way for future inquiry. First, this study selected companies that ultimately went bankrupt, precluding a direct comparison with a control group of non-bankrupt organizations under statutory audit. Future research should incorporate a control group to provide a more comprehensive understanding of GCO reporting practices across the entire spectrum of audited entities, including those that receive GCOs but successfully navigate financial distress through restructuring, refinancing, or other means.

Second, my reliance on publicly filed financial statements inherently excludes situations where audit opinions were not filed, particularly in the year before bankruptcy. While I acknowledge this data gap as a reflection of the very compliance issues I highlight, future research could explore alternative data sources or methodologies to gain insights into the reasons behind non-filings and the potential implications for auditor communication and accountability in these situations. Specifically, investigating instances where auditors allude to ongoing going concern assessments without a filed opinion is a promising avenue for future research, potentially using auditor interviews or regulatory inquiries (where feasible and permitted).

Third, my estimation of the societal costs of bankruptcy is based on insolvency claims at the time of filing and does not fully account for recovery rates or the long-term economic consequences. Future research could extend this analysis to incorporate recovery data and explore the broader macroeconomic impacts of corporate failures, further strengthening the societal relevance argument for effective going concern reporting.

Despite these limitations, this study provides valuable and timely empirical evidence on the state of auditor going concern reporting in the Netherlands. My findings contribute to the ongoing debate about audit quality, regulatory effectiveness, and auditor accountability, particularly in the context of non-listed SMEs. The persistent issues of non-compliance, limited disclosure, and Type II misclassifications revealed in my analysis call for urgent attention from regulators, professional bodies, and audit firms in the Netherlands to strengthen the going concern reporting ecosystem, enhance auditor accountability, and ultimately, better serve the public interest by providing stakeholders with more timely and reliable warnings of impending corporate failures. Moving forward, a renewed focus on enforcing filing regulations, promoting transparent management disclosures, and enhanced auditor vigilance in going concern assessments are crucial steps toward fostering a more accountable and resilient corporate environment in the Netherlands.