Skip to main content
Log in

Do going concern opinions provide incremental information to predict corporate defaults?

  • Published:
Review of Accounting Studies Aims and scope Submit manuscript

Abstract

Investors, regulators, and academics question the usefulness of going concern opinions (GCOs). We assess whether GCOs provide incremental information, relative to other predictors of corporate default. Our measure of incremental information is the additional predictive power that GCOs give to a default model. Using data from 1996 to 2015, initially we find no difference in predictive power between GCOs alone and a default model that includes financial ratios. However, there is an imperfect overlap between GCOs and other predictors. We show that GCOs increase the predictive power of several models that include ratios, market variables, probability of default estimates, and credit ratings. Using a model that includes ratios and market variables, GCOs increase the number of predicted defaults by 4.4%, without increasing Type II errors. Our findings suggest that GCOs summarize a complex set of conditions not captured by other predictors of default.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3

Similar content being viewed by others

Data availability

Data is obtainable from the sources described in the text and is available upon request.

Notes

  1. Auditing regulators worldwide have recently tried to increase the information content of the auditor’s report by requiring the discussion of key or critical audit matters (Gutierrez et al. 2018). However, GCOs focus on the client’s financial condition.

  2. The machine-based search criteria required the phrase “going concern*” in combination with the stem term “audit*” (e.g., audited, auditor, etc.) in the title or body of the article. For example, the press followed the General Motors case and Deloitte’s decision to issue a GCO for the company’s 2008 fiscal year-end (Johnson 2009).

  3. An entity must provide certain disclosures if “conditions or events raise substantial doubt about (the) entity’s ability to continue as a going concern.” The ASU applies to all entities and is effective for annual periods ending after December 15, 2016 (FASB 2014). In the Basis for Conclusions, ASU 2014–15 notes that in 2010 the FASB going-concern project considered proposing early-warning disclosures about going-concern uncertainties.

  4. The Hopwood et al. (1994) study examined 134 bankrupt firms and 160 nonbankrupt firms. The authors used a PD model based on six key financial ratios and firm sizes. The model was estimated using data from 1974 to 1981 and fitted to out-of-sample data from 1982 to1985. The two main takeaways are that inferences in previous studies were attributable to using (a) nonrepresentative samples, with 50% bankrupt and 50% nonbankrupt companies, and (b) mixed samples with both distressed and nondistressed companies. Hopwood et al.’s (1994) study is a widely used methodological reference in GCO research. However, the economic environment and the regulatory landscape for auditors have changed since its publication (e.g., the passage of the Private Securities Litigation Reform Act of 1995 and the Sarbanes-Oxley Act of 2002, along with the creation of the Public Company Accounting Oversight Board in 2002).

  5. We obtain data on corporate defaults from multiple sources, including the Center for Research in Security Prices (CRSP), Compustat, Bankruptcy.com, UCLA-LoPucki Bankruptcy Research, and the Credit Research Initiative (CRI). The Risk Management Institute at the National University of Singapore is a nonprofit organization that owns the proprietary CRI database. This database includes macroeconomic, financial, and default-related information. We complement these data with financial statement variables, stock prices, and auditor information from various sources, including a text-based search to identify GCOs before auditor opinions were available in Audit Analytics.

  6. The AUROC criterion has two useful aspects. (i) It summarizes a model’s true positive rates (i.e., correctly predicted defaults) and false positive rates (i.e., incorrectly predicted defaults) in a simple indicator that ranges between 0.5 and 1.0. (ii) Two models’ AUROCs can be compared using well-defined statistical tests, in sample and out of sample.

  7. The cited GCO accuracy studies use different sample periods and sources for GCOs and bankruptcy events. Geiger et al. (2005) and Geiger and Rama (2006) obtain GCOs from the SEC’s EDGAR database and bankruptcy events from the Bankruptcy Almanac. Feldmann and Read (2010) compile a list of companies that filed for bankruptcy from BankrutcyData.com and then locate audit reports for those companies in the SEC’s EDGAR database. Myers et al. (2014) use the Audit Analytics database to obtain relevant audit opinion data and the Bankruptcy Almanac to obtain bankruptcy events. Finally, Blay et al. (2016) use the Audit Analytics database to collect audit opinions and BankrutcyData.com to identify bankruptcy events.

  8. A related study by Hopwood et al. (1989) focuses on various forms of modified opinions and bankruptcy. They find that, among modified opinions, GCOs have statistical significance in a PD model that also includes financial ratios.

  9. This approach addresses a potential problem with previous default prediction methodologies. In prior methodologies, firms that disappear for reasons other than defaults are left out of the model, which can result in censoring biases (Duan et al. 2012).

  10. National University of Singapore, RMI, CRI database. Available at http://rmicri.org (accessed 27 January 2017).

  11. The Hopwood et al. (1994) model included ROA, LEVERAGE, CURRENT, CASH, two other ratios dividing current assets by total sales and total assets, and the natural logarithm of total sales as a proxy for size.

  12. There is no specific cutoff probability threshold prescribed by the auditing standards. An auditor is required to evaluate the going concern uncertainty after concluding that there is substantial doubt about the entity’s ability to continue as a going concern for a reasonable period.

  13. We use the Stata command rocreg to compare AUROCs. This command calculates standard errors based on bootstrap and does not assume that the ROCs are independent. This command is a modification of roccomp (Cleves 2002) that estimates standard errors following DeLong et al. (1988). As noted by Kim and Skinner (2012), although AUROC has advantages, a disadvantage is that it naturally increases as predictors are added to the model, in a manner analogous to unadjusted R-squares. There is no broadly accepted way of adjusting for this problem. However, our out-of-sample analyses, focusing on overall correct classification, are not affected by this problem. Finally, our cross-sectional analyses (Table 6) demonstrate that adding GCOs does not mechanically increase AUROC in all cases.

  14. As a robustness test, we confirm that our results hold when we include both distressed and nondistressed firms in a larger sample, finding that the combined model (Equation 5) has a statistically greater AUROC than any other model.

  15. We examined the non-overlapping GCO observations between our machine-based and the Audit Analytics (AA) classification. In total, there are 137 firm-years with a misclassification (out of 2181 GCOs available in AA). First, 37 were GCOs in our machine-based classification and not in the AA classification. Second, 100 were GCOs in the AA classification and not in our machine-based classification. Although these misclassifications constitute a small number of opinions, the biggest discrepancy resides in the second case, our machine-based algorithm failing to identify some GCOs. Out of the 100 misclassifications in the second case, we read 20 opinions and determined that the formatting of the text (e.g., page or line breaks in the middle of a sentence, unreadable text codes, etc.) in the 10-K filing prevented our algorithm from properly identifying the GCO. However, this issue would ultimately cause a reduction in GCOs’ predictive power by inducing noise in this independent variable. Also, this is a problem only for the period between 1996 and 1999, when we rely on our classification. Finally, our inferences are robust to limiting our sample to only the years covered by AA, starting with the year 2000.

  16. Our results are robust to the exclusion of CRI default data and to using only bankruptcies identified by the common data sources used in the literature.

  17. Our results are robust to using the 12-month CRI PD calculated at fiscal year-end, which aligns the end of the PD horizon with the end of the auditors’ GCO assessment horizon, and the six-month CRI PD calculated at the month end closest to the filing date.

  18. In general, our sample consists of relatively large firms with data in the Compustat, CRSP, and Audit Analytics databases, which are less likely to receive GCOs than the full population of U.S. listed firms (see Carson et al. 2013).

  19. Similarly, Blay et al. (2016) report that 1.79% of their financially distressed sample firms go bankrupt.

  20. Cook and Ramadas (2019) argue that precision-recall curves may be preferable to ROC when there are a small percentage of cases of interest (e.g., low incidence of default). We also examined whether adding the GCOi,t variable to the model with financial ratios, client size, and market variables increases the area under the precision-recall curve (Equations 4 and 5). We find that including the GCOi,t variable increases the area under the precision recall curve from 0.174 to 0.217, and the model with the GCOi,t variable has a higher precision along all the recall scale (zero to one).

  21. We thank Tyler Shumway (discussant) for suggesting this analysis and a percentage cutoff that mirrors the issuance rate of GCOs.

  22. Our inferences remain consistent when altering the cutoff threshold to the top 10, 20, and 30% of predicted probabilities. The pattern of increasing true-positive rates with the inclusion of GCOs remains similar at each cutoff, with the caveat that there are mechanically diminishing true-negative rates from the loosening of the cutoff threshold as illustrated in Figure 2.

  23. As explained in our methodology section, there is a well-defined statistical test that determines the change in predictive power when a new variable is added to a model. Unfortunately, we are unaware of a similar test that compares the marginal effect of adding a new variable to two models estimated in non-overlapping samples. This test implies “double differencing.” A limitation of our partition analyses is that they only provide inferences regarding whether GCOs add incremental information to both or only one partition. We find that GCOs add incremental predictive power in eight of our 12 partitions. In four partitions (i.e., level of distress, common versus uncommon distress conditions mentioned by the auditor, Big N, and high likelihood of opinion shopping), we conclude that the results do not support the notion that partitioning the sample along those criteria affects the incremental information of GCOs. Finally, we note that another limitation of these analyses is that they rely on subsamples of the data, where the low number of GCOs and defaults may reduce statistical power.

  24. It is difficult to empirically separate expected and unexpected GCOs. The determinants of GCO and default have a high degree of overlap (e.g., negative ROA and stock returns). By controlling for these joint determinants in a model that includes GCOs, the incremental AUROC arguably captures the orthogonal information component of the auditor’s opinion. This approach is similar to estimating a first-step model, predicting GCOs, and next including the residual (i.e., unexpected GCOs) in a second-step model of probability of default. Unfortunately, without strong exogenous variables to predict GCOs in the first step, the two-step procedure can actually result in worse inferences, due to the noise in two-step estimators.

  25. First, we model the likelihood of receiving a GCO in year t, where the independent variables are financial ratios, size, GCO in year t-1, and an indicator variable equal to one if the auditor is subsequently dismissed (DISMISS). Next, we estimate two probabilities, P1 and P0, based on the fitted model parameters for each firm-year observation, when DISMISS =1 and DISMISS = 0, respectively. Finally, we identify cases where the client seeks a favorable opinion: (1) the likelihood of receiving a GCO is higher when staying with the current auditor (P0 > P1), but the client switched auditors; and (2) the likelihood of receiving a GCO is lower when staying with the current auditor (P1 > P0), and the client did not switch.

  26. Other auditor incentives, beyond independence constraints, may also affect the results for the partitions based on client importance. For instance, the auditor may exert high effort for highly important clients, and this effort can result in incrementally informative opinions.

  27. Our main models (Equations 4 and 5) include overlapping determinants of the O-Score and the Altman Z-score, such as company size, working capital, current ratio, and ROA. Nevertheless, as a robustness test, we also examine whether GCOs provide incremental predictive ability to both the O-Score and the Altman Z-score, similar to our analyses based on the CRI company-level PD estimate. Including GCOs increases the AUROC of a model with the O-Score, Z-Score, and the non-overlapping control variables in the model with financial ratios and market variables. Existing evidence demonstrates that the CRI company-level PD estimate is a better predictor of performance than any other common prediction scores (Duan et al. 2012). Finally, a limitation of using all determinants is that they impose additional data constraints.

References

  • Akers, M., Maher, M., & Giacomino, D. (2003). Going-concern opinions: Broadening the expectations gap. The CPA Journal, 73(October), 38–42.

    Google Scholar 

  • Altman, E. I. (1968). Financial ratios, discriminant analysis and the prediction of corporate bankruptcy. The Journal of Finance, 23(4), 589–609.

    Google Scholar 

  • Altman, E. I. (1982). Accounting implications of failure prediction models. Journal of Accounting, Auditing and Finance, 6(1), 4–19.

    Google Scholar 

  • Altman, E. I., & McGough, T. P. (1974). Evaluation of a company as a going concern. Journal of Accountancy, 138(6), 50–57.

    Google Scholar 

  • Beaver, W. H. (1966). Financial ratios as predictors of failure. Journal of Accounting Research, 4, 71–111.

    Google Scholar 

  • Beaver, W. H. (1968). Market prices, financial ratios, and the prediction of failure. Journal of Accounting Research, 6(2), 179–192.

    Google Scholar 

  • Bharath, S. T., & Shumway, T. (2008). Forecasting default with the Merton distance to default model. Review of Financial Studies, 21(3), 1339–1369.

    Google Scholar 

  • Bills, K. L., Swanquist, Q. T., & Whited, R. L. (2015). Growing pains: Audit quality and office growth. Contemporary Accounting Research, 33(1), 288–313.

    Google Scholar 

  • Blay, A., Geiger, M., & North, D. (2011). The auditor’s going-concern opinion as a communication of risk. Auditing: A Journal of Practice and Theory, 30(2), 77–102.

    Google Scholar 

  • Blay, A., Moon, J., & Paterson, J. (2016). There’s no place like home: The influence of home-state going-concern reporting rates on going-concern opinion propensity and accuracy. Auditing: A Journal of Practice & Theory, 35(2), 23–51.

    Google Scholar 

  • Butler, M., Leone, A. J., & Willenborg, M. (2004). An empirical analysis of auditor reporting and its association with abnormal accruals. Journal of Accounting & Economics, 37(2), 139–165.

    Google Scholar 

  • Campbell, J. Y., Hilscher, J., & Szilagyi, J. (2008). In search of distress risk. Journal of Finance, 63(6), 2899–2939.

    Google Scholar 

  • Carson, E., Fargher, N., Geiger, M., Lennox, C., Raghunandan, K., & Willekens, M. (2013). Audit reporting for going-concern uncertainty: A research synthesis. Auditing: A Journal of Theory and Practice, 32 (supplement 1), 353–384.

  • Cassell, C. A., Hansen, J. C., Myers, L. A., & Seidel, T. A. (2017). Does the timing of auditor changes affect audit quality? Evidence from the initial year of the audit engagement. Journal of Accounting, Auditing & Finance, 1–27.

  • Chasan, E. (2009). Going-concern warnings increasingly less accurate. Reuters. https://www.reuters.com/article/goingconcern-study/going-concern-warnings-increasingly-less-accurate-idUSN2338896220090923. 28 Oct. 2019.

  • Chasan, E. (2012). Going concern opinions on life support with investors. The Wall Street Journal CFO Report. https://blogs.wsj.com/cfo/2012/09/12/going-concern-opinions-on-life-support-with-investors/. Accessed 28 Oct. 2019.

  • Chava, S., & Jarrow, R. A. (2004). Bankruptcy prediction with industry effects. Review of Finance, 8(4), 537–569.

    Google Scholar 

  • Cleves, M. A. (2002). From the help desk: Comparing areas under receiver operating characteristic curves from two or more probit or logit models. The Stata Journal, 2(3), 301–313.

    Google Scholar 

  • Cook, J., & Ramadas V. (2019). When to consult precision-recall curves. Working paper. Public Company Accounting Oversight Board.

  • Credit Research Initiative (2013). NUS-RMI Credit research initiative technical report. Version 2013: 2013 Update 2b. Global Credit Review, 3 (1), 77–167.

  • DeFond, M., Raghunandan, K., & Subramanyam, K. (2002). Do non-audit service fees impair auditor independence? Evidence from going concern audit opinions. Journal of Accounting Research, 40(4), 1247–1274.

    Google Scholar 

  • DeLong, E. R., DeLong, D. M., & Clarke-Pearson, D. L. (1988). Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach. Biometrics, 44(1), 837–845.

    Google Scholar 

  • Duan, J. C., Sun, J., & Wang, T. (2012). Multiperiod corporate default prediction: A forward intensity approach. Journal of Econometrics, 170(1), 191–209.

    Google Scholar 

  • Duan, J. C., & Van Laere, E. (2012). A public good approach to credit ratings –from concept to reality. Journal of Banking & Finance, 36(12), 3239–3247.

    Google Scholar 

  • Fawcett, T. (2006). An introduction to ROC analysis. Pattern Recognition Letters, 27(1), 861–874.

    Google Scholar 

  • Feldmann, D., & Read, W. (2010). Auditor conservatism after Enron. Auditing: A Journal of Practice & Theory, 29(1), 267–278.

    Google Scholar 

  • Financial Accounting Standards Board (2014). Disclosures about an entity’s ability to continue as a going concern. Accounting standards update (ASU) no. 2014–15. Norwalk, CT. https://www.fasb.org/jsp/FASB/Document_C/DocumentPage?cid=1176164329772&acceptedDisclaimer=true. Accessed 28 Oct. 2019.

  • Financial Reporting Council (2019). International standard on auditing (UK) 570 (revised September 2019). London, England: The Financial Reporting Council Limited. https://www.frc.org.uk/getattachment/13b19e6c-4d2c-425e-84f9-da8b6c1a19c9/ISA-UK-570-revised-September-2019-Full-Covers.pdf. Accessed 28 Oct. 2019.

  • Funcke, N. (2014). Credit rating changes and auditor reporting accuracy. Working paper. Erasmus University Rotterdam.

  • Geiger, M. A., Raghunandan, K., & Rama, D. V. (2005). Recent changes in the association between bankruptcies and prior audit opinion. Auditing: A Journal of Practice & Theory, 24(1), 21–35.

    Google Scholar 

  • Geiger, M. A., & Rama, D. V. (2006). Audit firm size and going-concern reporting accuracy. Accounting Horizons, 20(1), 1–17.

    Google Scholar 

  • Gerakos, J., Hahn, R., Kovrijnykh, A., & Zhou, F. (2017). Prediction versus inducement and the informational efficiency of going concern opinions. Working paper. University of Chicago and Arizona State University.

  • Gutierrez, E., Minutti-Meza, M., Tatum, K. W., & Vulcheva, M. (2018). Consequences of adopting an expanded auditor’s report in the United Kingdom. Review of Accounting Studies, 23(4), 1543–1587.

    Google Scholar 

  • Hillegeist, S. A., Keating, E. K., Cram, D. P., & Lundstedt, K. G. (2004). Assessing the probability of bankruptcy. Review of Accounting Studies, 9(1), 5–34.

    Google Scholar 

  • Hopwood, W., McKeown, J. C., & Mutchler, J. F. (1989). A test of the incremental explanatory power of opinions qualified for consistency and uncertainty. The Accounting Review, 64(1), 28–48.

    Google Scholar 

  • Hopwood, W., McKeown, J. C., & Mutchler, J. F. (1994). A reexamination of auditors versus model accuracy within the context of the going-concern opinion decision. Contemporary Accounting Research, 10(2), 409–431.

    Google Scholar 

  • Hosmer, D., & Lemeshow, S. (2000). Applied logistic regression (Second ed.). Hoboken: John Wiley & Sons.

    Google Scholar 

  • Janes, H., Longton, G., & Pepe, M. S. (2009). Accommodating covariates in ROC analysis. The Stata Journal, 9(1), 17–39.

    Google Scholar 

  • Jarrow, R. A. (2009). Credit risk models. Annual Review of Financial Economics, 1(1), 37–68.

    Google Scholar 

  • Johnson, S. (2009). The growing concern over “going concern.” CFO.com. https://www.cfo.com/accounting-tax/2009/03/the-growing-concern-over-going-concern/. Accessed 28 Oct. 2019.

  • Kim, I., & Skinner, D. (2012). Measuring securities litigation risk. Journal of Accounting and Economics, 53(1–2), 290–310.

    Google Scholar 

  • Koh, H. C. (1991). Model predictions and auditor assessments of going concern status. Accounting and Business Research, 21(84), 331–338.

    Google Scholar 

  • Larcker, D., & Zakolyukina, A. (2012). Detecting deceptive discussions in conference calls. Journal of Accounting Research, 50(2), 495–540.

    Google Scholar 

  • Lennox, C. (2000). Do companies successfully engage in opinion-shopping? Evidence from the UK. Journal of Accounting and Economics, 29(3), 321–337.

    Google Scholar 

  • Levitan, A. S., & Knoblett, J. A. (1985). Indicators of exceptions to the going concern assumption. Auditing: A Journal of Practice and Theory, 5(1), 26–39.

    Google Scholar 

  • Li, C. (2009). Does client importance affect auditor independence at the office level? Empirical evidence from going-concern opinions. Contemporary Accounting Research, 26(1), 201–230.

    Google Scholar 

  • Louwers, T. J., Messina, F. M., & Richard, M. D. (1999). The auditor’s going-concern disclosure as a self-fulfilling prophecy: A discrete-time survival analysis. Decision Sciences, 30(3), 805–824.

    Google Scholar 

  • Mayew, W. J., Sethuraman, M., & Venkatachalam, M. (2015). MD&a disclosure and the firm's ability to continue as a going concern. The Accounting Review, 90(4), 1621–1651.

    Google Scholar 

  • Menon, K., & Williams, D. (2010). Investor reaction to going concern audit reports. The Accounting Review, 85(6), 2075–2105.

    Google Scholar 

  • Merton, R. C. (1974). On the pricing of corporate debt: The risk structure of interest rates. Journal of Finance, 29, 449–470.

    Google Scholar 

  • Minutti-Meza, M. (2013). Does auditor industry specialization improve audit quality? Journal of Accounting Research, 51(4), 779–817.

    Google Scholar 

  • Mutchler, J. F., Hopwood, W., & McKeown, J. M. (1997). The influence of contrary information and mitigating factors on audit opinion decisions on bankrupt companies. Journal of Accounting Research, 35(2), 295–310.

    Google Scholar 

  • Myers, L. A., Schmidt, J. J., & Wilkins, M. S. (2014). An investigation of recent changes in going concern reporting decisions among big N and non-big N auditors. Review of Quantitative Finance and Accounting, 43(1), 155–172.

    Google Scholar 

  • Myers, L. A., Shipman, J. E., Swanquist, Q. T., & Whited, R. L. (2018). Measuring the market response to going concern modifications: The importance of disclosure timing. Review of Accounting Studies, 4 23 (4), 1512–1542.

  • Newton, N., Persellin, J., Wang, D., & Wilkins, M. (2016). Internal control opinion shopping and audit market competition. The Accounting Review, 91(2), 603–623.

    Google Scholar 

  • Ohlson, J. A. (1980). Financial ratios and the probabilistic prediction of bankruptcy. Journal of Accounting Research, 18(1), 109–131.

    Google Scholar 

  • Pepe, M. S., Longton, G., & Janes, H. (2009). Estimation and comparison of receiver operating characteristic curves. The Stata Journal, 9(1), 1–16.

    Google Scholar 

  • Public Company Accounting Oversight Board (2008). Staff audit practice alert no. 3 audit considerations in the current economic environment. Washington, D.C. https://pcaobus.org/Standards/QandA/12-05-2008_APA_3.pdf.

  • Public Company Accounting Oversight Board (2011). Staff audit practice alert no. 9 assessing and responding to risk in the current economic environment. Washington, D.C. https://pcaobus.org/Standards/QandA/12-06-2011_SAPA_9.pdf. Accessed 28 Oct. 2019.

  • Public Company Accounting Oversight Board (2012). Investor sub advisory group. Going concern considerations and recommendations. Washington, D.C. https://pcaobus.org/News/Events/Documents/03282012_IAGMeeting/Going_Concern_Considerations_and_Recommendations.pdf. Accessed 28 Oct. 2019.

  • Public Company Accounting Oversight Board (2014). Staff audit practice alert no. 13 matters related to the auditor’s consideration of a company’s ability to continue as a going concern. Washington, D.C. https://pcaobus.org/Standards/QandA/09222014_SAPA_13.pdf. Accessed 28 Oct. 2019.

  • Public Company Accounting Oversight Board (2017). Consideration of an Entity’s ability to continue as a going concern. Washington, D.C. https://pcaobus.org/Standards/Auditing/Pages/AS2415.aspx. Accessed 28 Oct. 2019.

  • Shumway, T. (2001). Forecasting bankruptcy more accurately: A simple hazard model. Journal of Business, 74(1), 101–124.

    Google Scholar 

  • Vassalou, M., & Xing, Y. (2004). Default risk in equity returns. Journal of Finance, 59(2), 831–868.

    Google Scholar 

  • Vanstraelen, A. (2003). Going-concern opinions, auditor switching, and the self-fulfilling prophecy effect examined in the regulatory context of Belgium. Journal of Accounting, Auditing and Finance, 18(2), 231–254.

    Google Scholar 

  • Witten, I. H., & Frank, E. (2005). Data Mining: Practical Machine Learning Tools and Techniques. Morgan Kaufmann series in data management systems, Second edition. Morgan Kaufmann.

  • Zmijewski, M. E. (1984). Methodological issues related to the estimation of financial distress prediction models. Journal of Accounting Research, 22, 59–82.

    Google Scholar 

Download references

Acknowledgements

The authors are grateful for helpful comments from Peter Easton (editor), two anonymous reviewers, Pietro Bianchi, Khrystyna Bochkay, Lauren Cunningham, David Hay (discussant), Mark Maffett, Linda Myers, Dhananjay Nanda, Sundaresh Ramnath, Tyler Shumway (discussant), Eric Weisbrod, Michael Willenborg, Peter Wysocki, PCAOB staff, and seminar participants at the AAA Auditing Section Midyear Meeting 2016, University of Chicago, Florida Accounting Symposium 2016, University of Miami, University of Minnesota Empirical Accounting Research Conference 2016, Nanyang Technical University, National University of Singapore, PCAOB/JAR Conference on Auditing and Capital Markets 2017, and University of Toronto. The authors also thank Yamin Hao and Taylor Wiesen for their excellent research assistance. The authors thank the Risk Management Institute at the National University of Singapore for providing the CRI default data.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Miguel Minutti-Meza.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix 1

Appendix 1

1.1 Variable Definitions

Variable

 

Definition

Source

DEFAULT i,t + 1

=

indicator variable equal to one if company i has a default event from the filing date of the annual report for fiscal year t to the subsequent fiscal year end in t + 1 and 0 otherwise.

CRI, CRSP, Compustat, Bankruptcy.com, UCLA-LoPucki Bankruptcy Research Database

GCO i,t

=

indicator variable equal to one if a company i has a GCO at fiscal year-end t and 0 otherwise.

Audit Analytics and text mined SEC 10-K Filings

CONT_PD i,t

=

Estimated PD for the subsequent 12 months for firm i calculated at the closest month end before the signature date of the audit opinion for fiscal year t.

CRI database

REL_MKTCP i,t

=

Log (Firm Market Capitalization [PRC*SHROUT] / Index Market Capitalization [TOTVAL]).

CRSP

EX_RET i,t

=

Cumulative Firm Returns – Cumulative Market Returns over the 12 months leading up to the filing date for year t.

CRSP

SIGMA i,t

=

Standard deviation of the residuals from a monthly return model of firm returns on market returns for the 12 months leading up to the filing date for year t.

CRSP

ROA i,t

=

Net Income [NI] / Total Assets [AT].

Compustat North America

LEVERAGE i,t

=

Total Liabilities [LT] / Total Assets [AT].

Compustat North America

WCAP i,t

=

(Current Assets [ACT] – Current Liabilities [LCT]) / Total Assets [AT].

Compustat North America

CURRENT i,t

=

Current Assets [ACT] / Current Liabilities [LCT].

Compustat North America

CASH i,t

=

Cash and Cash Equivalents[CHE] / Total Assets [AT].

Compustat North America

CFO i,t

=

Cash Flow from Operating Activities [OANCF] / Total Assets [AT].

Compustat North America

SIZE i,t

=

Log(Total Assets[AT]).

Compustat North America

NEGEQUITY i,t

=

Indicator variable equal to one if Total Liabilities [LT] exceed Total Assets [AT] and 0 otherwise.

Compustat North America

BIGN i,t

=

Indicator variable equal to one if the company has a Big N auditor and 0 otherwise.

Audit Analytics and Compustat (pre-2000)

RATING i,t

=

Categorical variable representing the credit rating [CR_LVL] of the firm at fiscal year-end t, ranging from 1 (AAA) to 10 (D).

Compustat North America

DOWN i,t

=

Indicator variable equal to one if the credit rating [CR_LVL] for firm i has been downgraded during the 12 months of fiscal year t.

Compustat North America

DOWN_LVL i,t

=

Count variable indicating the number of levels in credit ratings [CR_LVL] that firm i has been downgraded during the 12 months of fiscal year t.

Compustat North America

INVST_CH i,t

=

Indicator variable equal to one if the credit rating [CR_LVL] for firm i has been downgraded from investment grade to non-investment grade during the 12 months of fiscal year t.

Compustat North America

  1. Relevant Compustat and CRSP variable names in brackets

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Gutierrez, E., Krupa, J., Minutti-Meza, M. et al. Do going concern opinions provide incremental information to predict corporate defaults?. Rev Account Stud 25, 1344–1381 (2020). https://doi.org/10.1007/s11142-020-09544-x

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11142-020-09544-x

Keywords

Navigation