Bad news and bank performance during the 2008 financial crisis

Abstract

The article investigates market reaction to negative reports published by analysts and auditors for a sample of investment, commercial and savings banks during the 2008 financial crisis and compares the results to noncrisis periods. The results show that during 2008, analysts’ downgrades and underperformance reports resulted in stronger negative returns than during noncrisis periods and that investment banks experienced the worst stock price declines. The market reaction to auditors’ issues and going concern flags is different during the crisis as well. In noncrisis periods no reaction to auditors’ bad news is reported, while during the crisis there is a negative and significant reaction for investment banks only. Overall, the type of bank, investment versus commercial, significantly contributes to explaining the variability in returns during the financial crisis.

Keywords:

• financial crisis
• auditors
• analysts
• bank performance

JEL Classification:

• G01
• G21

Notes

¹  Issue flag (represents a red flag or compliance issue) has been discontinued from Audit Analytics as of 11 April 2011. The only field that currently represents bad news is GC.

²  One of the criticisms of auditor opinions is that they are not issued until later. For example, 2008 audit reports were not issued until 2009. This, however, does not represent a problem in this analysis as we are assessing the reaction to bad news issued during the crisis period and how the reaction compares to noncrisis periods. We measure the market reaction to bad news when it reaches the market in general (regardless of the original date the issue is documented) rather than to specific bad news about particular companies.

³  The sample size for going concern is very small, only 28 total flags, most of them in 2008. Therefore, the majority of analysis for the auditor portion is performed based on the issue field that represents a red flag or potential compliance issue.

⁴  We also use alternative specifications found in literature, such as (−2, +2) and (−1, +7) with no differences in results.

⁵  Alternative structures, such as day (0), (−2, +2) and (−1, +7) have also been examined with no significant differences.

⁶  These results correspond to the equally weighted portfolio. For a value weighted portfolio, the declines are 1.98% for the 3-day window and 1.26% for day (0). From here on, the first return is calculated based on the equally weighted and (second) on the value weighted portfolio.

⁷  There are large differences between the equally weighted and value-weighted portfolios for savings banks due to the domination of the returns by a few large banks. That is one reason why I separated this group from the commercial banks.

⁸  The sample size was very small for going concern before the crisis and is therefore omitted.

⁹  The sample size for the 2008 GC is 28 firms.

¹⁰ Although I present the results for the both equally weighted and value-weighted portfolios, due to the disproportionate sizes of the banks, the value-weighed portfolio should be used for the analysis.

¹¹  In noncrisis periods, 2000 to 2007 and 2009 to 2010, there are no abnormal returns to auditors’ issues that require a cross-sectional explanation.