Listen to the podcast:
Audit reports — and the requirement that public companies file audited financial statements — are a cornerstone of modern financial reporting. The Securities Exchange Act of 1934 requires that all public companies disclose audited financial statements and their findings. But could these audits, which are designed to protect shareholders, actually provide a motivation for insider trading?
In the paper “Audit Process, Private Information, and Insider Trading,” a team of researchers from Wharton and Indiana University examine whether corporate insiders trade based on foreknowledge of audit outcomes. They found a spike in trading activity around the date at which audit findings are conveyed to the Board of Directors but not the public. The paper was recently awarded the ‘Best Academic Paper’ at the John L. Weinberg Center for Corporate Governance Annual Symposium.
One of the paper’s co-authors, Wharton accounting professor Daniel Taylor, joined Knowledge@Wharton for a deep dive into their research, which has implications for corporate boards and regulators.
An edited transcript of the conversation follows.
Knowledge@Wharton: Can you give us a bird’s eye view of your research and what you found?
Daniel Taylor: Sure. So big picture, my research in general uses big data and analytics tools to discern patterns and detect opportunistic reporting and trading by officers and directors. This particular paper was based on a large-scale data analysis of all Form 4’s filed by officers and directors of all publicly traded companies. And for those who aren’t aware, a Form 4 is where an officer and director has to publicly disclose the trades that they make in their firms.
Now a couple of caveats. Just to be clear, because we are using Big Data one limitation of the analysis is that while we can identify suspicious behavior and raise red flags about certain individuals and companies, we can’t really speak to legality. So when I use the words ‘insider trading,’ I am talking about trading of corporate insiders — managers and directors — not the illegal insider trading that you might see in hedge funds and whatnot.
In this particular case, we are going to get into a gray area in which it really looks like it might be illegal, but because we’re analyzing it at such a high level there could be extenuating circumstances. There’s a specific context that would need to be taken into account in each particular case. This highlights the tradeoff when using Big Data: We can see recurring patterns but we don’t do a deep dive on any one company.
“Do insiders trade based on their foreknowledge of the audit outcome?”
This particular paper examines whether corporate insiders trade on information about audit outcomes before those audit outcomes are publicly disclosed. We took a ‘Sherlock Holmes’ approach to the data — we have data on audit outcomes, who performed the audit, who is trading, and when they are trading. We then correlate these data … to see what patterns are in the data and what they suggest.
One important thing to keep in the back of your mind to provide some context on the research is that a lot of firms have what is known as ‘insider trading blackout windows.’ Those blackout windows govern when the company allows insiders to trade. Those blackout windows are not required by the SEC, nor are they enforced by market regulators. The company can choose whether to have them and also whether to enforce them.
In some previous work that I have done, we did a survey of common blackout windows and found that most of the blackout windows are written around earnings announcements. So a blackout window might say the insider cannot trade, for example, two months before the earnings announcement through two days after the earnings announcement.
But [since then], there have been some big changes. Those of you who follow markets know that the Sarbanes-Oxley [Act of 2002, which established sweeping auditing and financial regulations on public companies,] was a big change. Before Sarbanes-Oxley, most audits actually came out at the same time as the earnings announcement. For example, in 2002, only 25% of firms released their earnings announcement before the completion of the audit.
But now fast forward to 2016, and 60% of companies release their earnings announcement before the audit. And that is because after Sarbanes-Oxley, audits are taking much longer. So now it is very common for companies to release earnings with the audit results not being finalized until many, many days later. Consequently, those audit results will be outside of the earnings announcement blackout window — when the firm allows insiders to actually trade.
… Therefore, a natural question is whether that longer audit is potentially providing managers with private information that they are exploiting. … Managers and directors know what the audit findings are before the public, and now we have an institutional setting where the blackout window may in fact be open at some companies, in our sample. So the question our research seeks to answer is rather simple: Do insiders trade based on their foreknowledge of the audit outcome?
Knowledge@Wharton: For context, what are some of the rules around insider trading currently?
Taylor: It is illegal to trade on material, non-public information. That applies to everyone: officers, directors, hedge fund managers, etc. But officers and directors also have something known as a fiduciary duty. That fiduciary duty [would necessitate that] they adhere to a rule known informally as the ‘disclose or abstain’ rule. So if the director or the officer has material, non-public information, their duty compels them to either disclose that information to shareholders or abstain from trading. That is what governs the legal trading behavior of officers and directors.
“Our research shows that there is a spike in trading around this audit report date.”
In our analysis, we focus on the audit report date. The PCAOB (Public Company Accounting Oversight Board) basically requires the board of directors to be briefed on or around that audit report date. That audit report doesn’t come out to the public until many days later when the 10-K is filed [after the earnings report], giving insiders an opportunity to profit from any crucial information in the audit report that would have a material effect on stock prices [such as whether the firm is a going concern, weak internal controls or deficient accounting practices].
Our research shows that there is a spike in trading around this audit report date. While we can’t say it is illegal because of the caveats that I mentioned earlier, we identify a certain number of firms and individuals where it is as close as a large sample analysis can come to drawing such an inference.
Knowledge@Wharton: Let’s dig deeper into your paper, and talk about the step-by-step process of what happens in an audit of a public company.
Taylor: I think most people think that the audit is something that is very, very short. It takes a couple of months, and then it is done, there is a document or a certification that the audit is done with the audit findings. And that is really a myth because a typical audit is really a year-round process.
Here’s the timeline: Audit planning and interim procedures occur during the fiscal year that test the company’s internal controls over financial reporting and accounting transactions. Then there will be something called ‘year-end field work,’ where there will be auditors out in the field doing their work around the annual earnings announcement. If the audit is progressing as it should, there will be a status update to the board shortly before the company announces earnings. It should say ‘Here are some outstanding issues; here is where we’re at.’
Then the company will announce earnings. For 60% of companies in recent years, the audit is not complete by that time. But the process is completed somewhere between the earnings announcement and the 10-K filing date with the preparation of the audit report. PCAOB auditing standard 1301 requires that the board is briefed close to the date that the audit report is finalized. That is known as the ‘audit report date’ and signifies the completion of the audit.
That date is not known to the public, so you can think of it as a board briefing date. The firm subsequently discloses the audit report in the 10-K filing. So there are really four dates at work here: the fiscal year end, the earnings announcement, the audit report date, and the 10-K filing date. Remember that the audit report date isn’t going to be known to the market until the audit report is filed as a part of the 10-K. … We looked at trades between the earnings announcement and the 10-K, to see whether there are changes in trading behavior in that window, and whether that correlates with the date of the audit report and the contents of the audit report.
As for the contents of the audit report, it can issue what we call a clean opinion, which means a standard boilerplate opinion with no changes. Everything looks great, internal controls and the accounting practices conform with the standards. Or, auditors can issue something known as a modified opinion. It is too strong to say that a modified opinion means the company failed the audit [but it does flag some problems.]
Within the modified opinion there is a lot of explanatory language, and the explanatory language that could flag the firm as a going concern — meaning auditors are concerned about the firm’s continued ability to operate, to be in business. There could be concern about material weaknesses in internal controls over financial reporting.
“We are finding relatively more trading when the audit report contains a modified opinion, which is not a clean bill of health for the company.”
The auditor could flag certain issues of concern having to deal with, say, past stock compensation expense; or option back-dating; or a restatement; or that a new accounting standard has had a material effect on the company’s operations. What is interesting about these reports is that they are all, in some sense, boilerplate. It is only a few key sentences that really provide all of the information — such as a going concern or a material weakness.
Knowledge@Wharton: I want to hone in on something you mentioned, which was the timing of the trades that you looked at in your research. There were some very specific reasons why you chose that window. Can you explain that a little bit?
Taylor: We’re looking at the trades of all officers and directors that are filed on Form 4 with the SEC. There are a couple of caveats. We are going to end up restricting the sample to a very narrow set of firms. … We are going to drill down to those firms where the audit report comes after the earnings announcement and more than 10 days before the 10-K filing. So we looked at a very narrow set of firms, 1,379 firms in all.
For these firms, the audit report on average is finalized 20 days after the earnings announcement and 20 days before the 10-K. So if a firm announces earnings and some 40 to 45 days later it files a 10-K, right smack in the middle of that window is the audit report date. We looked at trades five days before and five days after the audit report date. These trades are likely not about earnings because that report came out 20 days ago. They also are not likely about the 10-K per se … because the 10-K will be filed in 20 days.
We focused on this wide 40-day window between the earnings announcement and the 10-K and we conducted an ‘event study.’ We looked for a sharp discontinuity, or break, in the trading behavior of officers and directors in the five-day window around the audit report date. We chose that window intentionally because we want to try and rule out as many alternatives as possible. … By focusing on … this narrow five-day window, it makes it less likely that there is an alternative explanation [for the trading spike] because there is no other systematic corporate event occurring at the same date as the audit report.
There also aren’t any systematic public news events going on that could potentially generate these trades. We focus very narrowly on this sample of firms and this particular date because we think this is where we can best identify opportunistic behavior in conjunction with the audit. That is not to suggest it is not going on elsewhere, but we think we feel most comfortable saying it is definitely going on in this narrow sample.
Knowledge@Wharton: What kinds of securities did you actually look at in this insider trading activity?
Taylor: We looked at the open market purchases and sales of [stock by] officers and directors. So a stock option grant or a restricted stock grant would not be considered. This has to be the executive actually selling shares on the open market, or conversely purchasing shares on the open market. … They are trading with the investing public more generally.
Knowledge@Wharton: Why did you exclude options, stock options, stock grants, etc.?
Taylor: With options and grants, we didn’t want this to be a compensation story.
… What we documented … is a flat line [of trading activity] leading up to the audit report date. But in the five days around the audit report date, there is this pronounced spike, and then after that it goes back down — just like how an electrocardiogram would detect a heartbeat. So it is less likely [that people are trading because they need] liquidity, because why would you need to sell your shares to buy a house … on the same date in which you are briefed by the auditor on the outcome? So that is what we used to rule out other, more benevolent uses for funds.
… [And] not only are we finding a pronounced increase in trading around the audit report date, we are finding relatively more trading when the audit report contains a modified opinion, which is not a clean bill of health for the company. The opinion might mention a going concern, or material weakness, it might have explanatory language around a particular accounting issue.
“It is not the officers that are trading after the audit report — directors are the ones trading.”
… The other interesting thing is we do not observe this phenomenon in the other quarters: quarter one, two and three. [Only the fourth.] In the first three quarters, they have blackout windows and there are still earnings announcements but you don’t observe the same thing. Because there is no audit report date for quarters one, two, and three.
We have tried to throw a bunch of explanations at this phenomenon that most reasonable people would agree would be predictions of alternative explanations and see if we could rule it out. … The fact that the spike varies with whether the company gets a clean bill of health from the auditor is quite astounding to us, because as academics we always approach something from a very skeptical or cynical perspective.
The other thing I should mention is that because we have the Form 4s we can drill down on individuals who are trading and what are their roles in the company. So for example, is it the directors who are trading? Is it the officers that are trading? Is it the audit committee members who are trading? Is it the non-audit committee members who are trading?
The data suggests that it is not the officers that are trading after the audit report — directors are the ones trading on or shortly after the audit report, and officers tend to trade before the audit report, which in some sense is consistent with the timing of when we might think certain directors might learn some information.
Officers and audit committee members might be aware of the audit findings before the entire board, but the board members that are not on the audit committee, who are typically not financial experts, are more likely to learn about the audit for the first time around the audit report date.
Knowledge@Wharton: You also said in your paper that looking at insider trading activity based on audit reports is particularly critical this year because of a change in auditing standards. Can you talk about that a little bit?
Taylor: It’s one reason why I think the paper is very timely. The auditing standards go into effect at the close of 2019 and 2020 that require auditors to include a discussion of what is known as critical audit matters in the audit report.
Critical audit matters are matters that have been communicated to the audit committee, and to the board by the auditor, and are related to disclosures that have a material effect on the financial statements and that involve especially challenging and subjective auditor judgment. … There is an incredible amount of subjective judgment that goes into whether an accounting treatment is permissible, or whether it crosses a materiality threshold that would require a restatement or a change in disclosure.
… Former PCAOB chair James Doty gave a speech basically saying that the new audit report is going to give the investors the information that they have been asking for … that is material to their decision-making. Now that information is already being communicated to the board, but under the old regime it is just not being communicated to investors. Now it will — so there will be less of an information gap when it comes to the audit between the board and investors.
“This is one of the first papers to look at trading around what we call an internal information event.”
There certainly will still be an information gap, but maybe a little less. What that will mean is that the audit report will be highly engagement-specific. There will be less boilerplate involved because it will be more involved with what the issues were at this particular company, or that is at least the hope. No one knows. We haven’t seen these come out in great numbers yet.
It is a double-edged sword. If the new auditing standard succeeds at making the audit report more meaningful, and it therefore causes stock prices to change more, the audit report becomes a more material disclosure. Perversely, that increases the incentive for individuals to front-run the disclosure of the audit report.
Knowledge@Wharton: What message do you hope to send to company boards, to regulators, and academics regarding your findings?
Taylor: One of the key implications is that we need to limit the trading of personnel involved with the audit. Now again, I want to underscore that the study is primarily concerned with firms where the audit is completed after the earnings announcement. For such firms there are two [solutions]: First, delay announcing earnings until the 10-K filing — so the earnings announcement, the audit report, and the 10-K are all released on the same day. The second option is to alter the terms of the firm’s insider trading policy. … You might want to lengthen the blackout window to extend through the disclosure of the 10-K.
Knowledge@Wharton: How is your research different from prior work in this area?
Taylor: There is a lot of academic work on insider trading. … But what is more surprising to us is that there is very little around what we will call internal information events. So for example, there is a lot of work on insider trading around earnings announcements, and lots of work around insider trading around equity offerings and insider trading around share buybacks, and dividend announcements — all of those are public events.
But this is one of the first papers to look at trading around what we call an internal information event, where the event isn’t known to the market at the time that it occurred. Another example would be a board meeting. Board meeting dates do not have to be publicly disclosed. And so we are in the process right now of trying to acquire data sets on board meeting dates, and then examine trading around those board meeting dates.
Knowledge@Wharton: How will you follow up this research?
Taylor: One thing that we are thinking about doing is looking at trading around the audit report to see whether it predicts future restatements and fraud cases. That is where the audit would be the most material. Imagine being on the board when you learn that the auditor has detected some irregularity with the inventory, or detected some irregularity with a special purpose entity.
You may not be engaged in the fraud yourself, but this will be your first chance to see the auditor’s concern and expression about these instances of fraud or of major accounting issues. And you may be tempted to trade on that information. So what we are thinking about doing is looking at trades by the board around the briefing date, and to see if that predicts, using historical data, whether there will be a fraud or a restatement. The initial evidence is that it looks like it will be predictive.