An interview with Mark Flannery, SEC Chief Economist and Director of DERA
Dr. Mark Flannery is the Chief Economist at the SEC and the Director of the SEC's Division of Economic and Risk Analysis (DERA). He joined the SEC in September 2014. Dr. Flannery is on leave as a distinguished scholar in finance at the University of Florida and was a senior advisor at the Treasury Department's Office of Financial Research from 2011 until 2014. Dimensions spoke with him recently by telephone.
You have been a highly respected finance professor at the University of Florida, you have authored many scholarly articles, and you have served as an advisor or a visiting scholar in prestigious places. What led you to accept your position at the SEC?
Well, there are really two elements to that. First, I had been a finance professor for a long time. I knew I could do that, I knew I could continue doing that, and I knew I would enjoy it. But I was looking for a different challenge that would get me out of my comfort zone. I succeeded admirably in taking this job! Why the SEC? DERA, the Division of Economic and Risk Analysis, is in a very strong position relative to what it was three, four, or five years ago. It is involved in a great deal of Commission rulemaking. It is involved in giving a lot of economic advice about SEC rules and areas of possible risk exposure. And these are essentially the first years in which we have been able to introduce economic analysis into the entire fabric of the SEC. DERA was a much smaller operation previously. This seemed like a really exciting opportunity to get involved in introducing economic analysis into this regulatory space.
You probably had an impression of the SEC before you joined the agency in September 2014. Now that you are on the inside, how does the reality of the SEC differ from your impression?
“The SEC is much more multidimensional than I had thought it was going to be.”
Has anything about the SEC surprised you? Oh, many things. I had previously known a lot about bank regulators and relatively little about the SEC. What I have been struck by is the complexity that lives inside this building - the number of issues, the number of perspectives, and the number of responsibilities for making the markets work as well as they do in this country. It is much more multidimensional than I had thought it was going to be, and it has been a real challenge to learn about the issues that are so current here.
Now you are both the Chief Economist and the Director of DERA, two very challenging positions. How do you perform both roles?
Really they are two different names for what is much the same thing. “Chief Economist” is an outward-facing title and “Director of DERA” is an inward-facing title (because people on the outside do not really know what DERA is). The Chief Economist's role involves providing economic oversight to the very talented staff here at the Commission. Recruiting is also important, and I try to draw in talented folks and to identify the issues to which we should be applying their talents most directly. The work of the Director of DERA is more internally focused - it is more a management job. That means dealing with people inside the Division and relating their work and our work to other divisions and other decision makers inside the SEC. They are highly complementary positions with slightly different focuses.
In a recent speech, you discussed CIRA, the SEC's Corporate Issuer Risk Assessment program, formerly known as the Accounting Quality Model (AQM). How has the AQM been enhanced for the CIRA program?
“CIRA allows SEC staff to compare any or all of 100 metrics to different firms within a peer group.”
CIRA is essentially the Accounting Quality Model on steroids. Instead of looking at just accounting anomalies or accruals, which was the main focus of the AQM, CIRA looks at all dimensions of financial data. We have about 100 different metrics that describe firms. CIRA allows SEC staff to compare any or all of those 100 metrics to different firms within a peer group. Essentially, it is a dashboard that allows us to look for possible anomalies that may warrant further investigation. So we have gone far beyond anomalous accounting information, or accounting quality, to look at a bigger group of variables and to give more flexibility to the SEC staff who use the tool.
What are some of CIRA's 100 metrics?
We have kept the accruals concept from the original AQM. But we also look at performance: What is a firm's stock price or its total shareholder return or its accounting variables (say, for example, growth and earnings) relative to others in a peer group? We have some information about governance measures - for example, information about the board. We have dozens of standard financial ratios: return on assets, return on equity, etc. The goal is to have, as a possible metric, anything that you might see in an accounting textbook that would be useful for evaluating a firm in general.
How much of the data analyzed with CIRA comes from XBRL filings?
Right now, the data come primarily from commercial databases, but some of the XBRL variables that are reported to us also go into the tool. I want to emphasize that generally we use XBRL data where it is available today, and we are very interested in expanding the availability and use of XBRL data. In the context of CIRA, as we get more XBRL data, and as we get a longer time series in XBRL data reports, we will incorporate more and more of that information directly into CIRA. With regard to the use of XBRL more generally, one of the important things about our XBRL filings so far is that they are broader - they cover more firms than are covered by the commercial database-providers. Commercial database providers do not provide as much information about small firms, which may be of less interest to their customers. With XBRL reporting, which includes data from small firms, we have a larger universe of firms, and that is going to be one of the benefits of expanding XBRL use in the CIRA dashboard.
Some business journalists used the nickname “RoboCop” to refer to the AQM, the predecessor of the CIRA program. You have said in a few speeches that you do not like this name. Why not?
It is catchy and implies an automated monitoring of SEC filings and financial risks. CIRA is a tool. It is useful only if it is being applied by somebody who knows how to use the tool. CIRA can identify potential outliers or potential unusual situations, but until expert staff has looked at the overall situation, we do not have any information about whether the appearance of something anomalous is actually the source of a problem. So the tool itself is useless without the expertise of the staff who uses the tool.
DERA's Office of Interactive Data has been renamed the Office of Structured Disclosure. Does this reflect a new attitude at the SEC about the importance of structured data and structured disclosure?
I do not think it reflects a new attitude but rather an effort to bring out the expertise that is in our Office of Structured Disclosure, let it stand on its own, and let it be visible outside DERA so that staff inside the SEC as well as registrants who need information about structured data have a place to call and a place to go. The Office of Structured Disclosure has people who are experts in various disclosure mechanisms - XBRL being one, but not the only one. Every time we make a new rule or decide to ask for some new information, the staff in DERA, including in the Office of Structured Disclosure, will be involved in considering what information is required, and how it should be reported. The office also has a helpline, “Ask OID,” where people can submit questions about a particular XBRL reporting requirement or about tagging a particular kind of field. In sum, we picked out a part of DERA that had become large enough to stand on its own and be recognized as a separate specialization.
Companies and Congress are looking for signs that the SEC is actually using XBRL data. Can you provide examples of ways in which the SEC has successfully used XBRL in its review of filings and in other activities?
“XBRL expands the number of smaller firms beyond those normally included in a commercial database. Compared to the situation before XBRL, the number of firms whose financial statements we can provide has increased by about one-third.”
Certainly. One of the most visible manifestations of our interest in and use of XBRL is that every quarter we publish sets of financial statements, available on the web for downloading in a structured file (basically a flat file). These data sets are created from tagged data provided in Forms 10-K and 10-Q. These come out very quickly after the end of each quarter, illustrating how XBRL makes it easy and quick for us to put out information. This information is important not because it replicates what the commercial databases have, but because it expands the number of smaller firms beyond those normally included in a commercial database. Compared to the situation before the XBRL rules were adopted, the number of firms whose financial statements we can provide has increased by about one-third. Another useful thing for us is that footnote information is now tagged. That gives us an opportunity to develop a richer set of data than is available in the face financials - by looking at the footnote information and picking out sets of topics that are treated in the footnotes. As we go forward, XBRL filings will increase in number. We will have more and more years of data. That will naturally increase the use of XBRL data.
Do you expect an increase in the amount of XBRL data used in the CIRA program?
Does the SEC plan to use XBRL data in place of information currently purchased from commercial aggregators?
Every year, we have a process for evaluating the commercial databases we need. The question always is: What is available that we do not have, and are commercial offerings worth the price? The more information we can generate and access from within the SEC, the fewer dollars we will need to spend on external data. I do not have a list of data I think will go away, but I do expect that the XBRL availability will tend to displace some of the commercial data that we have been buying.
Is the SEC going to issue more staff observations, updates to filing practices, or "Dear CFO" letters? Has there been any discussion about using comment letters to get companies to improve the quality of their XBRL use?
“Analysis of 2014 filings is under way within DERA. I would not be at all surprised to see further information about XBRL issues.”
The quality issue is being addressed. For example, DERA has produced a report on the use of custom tags. The more custom tags a filer uses in an XBRL filing, the more difficult it is to compare that particular filing to filings that are using standard tags. DERA presented a report about the incidence of custom tags and, it turns out, they were particularly prominent among filings from smaller companies that were newer to the process. DERA also analyzes XBRL data to help the Division of Corporation Finance in their interactions with registrants. Last summer, Corporation Finance sent out a letter that specifically addressed certain issues about required calculations inside the XBRL data.
Do you expect to issue more staff observations?
Analysis of 2014 filings is under way within DERA, and we will provide that information to the Division of Corporation Finance. I would not be at all surprised to see further information about XBRL issues.
When will DERA's report on the 2014 filings be issued?
During this summer? At the moment, that is the expectation.
In a speech, you mentioned that the SEC has seen improvement in the quality of XBRL filings by large and mid-sized companies. What is the SEC doing to help small companies improve the quality of their filings?
There are more XBRL service-providers than there were five years ago, when the large firms started reporting. One thing we are doing is helping the small companies by increasing their awareness of these service-providers. Inline XBRL may also simplify the preparation of financial statements and could help improve data quality.
What is inline XBRL? What are its benefits?
Instead of having to prepare financial statements twice - once in a version for people to read (HTML) and once in a version that is tagged (XBRL) - filers could use inline XBRL to produce one document that is cheaper to produce and easier to check for potential tagging errors. Whether inline XBRL is adopted is a Commission decision, but DERA is considering the potential benefits of providing inline XBRL as an option for filers. Since cost savings may be more important to small companies, inline XBRL may be of more significance to them. We are also developing a prototype viewer for reading inline XBRL so that we can translate the gobbledegook into something that people can read and understand.
Given the reporting latitude that GAAP provides for financial disclosures, do you see a way in which XBRL can make financial statements more comparable?
The issue of comparability is not just about XBRL. It is essentially about the application of GAAP. I remember speaking with an accounting professor at the FASB, and I asked, in my naivety, “Why don't you just let everybody tell their own financial story the way they like?” And she said, “Because a very important principle of GAAP is that we try to constrain the way companies tell their financial stories so that we can compare across companies.” XBRL does not drive comparability. Rather, XBRL conveys GAAP through a taxonomy. And to the extent that GAAP is good at allowing cross-company comparisons, that is what will allow for comparability. XBRL will be the medium through which comparability happens, but it will not be the cause of comparability.
The SEC recently proposed pay-versus performance rules for proxy filings that include an XBRL-tagging requirement. Are there plans to adopt XBRL for other parts of the proxy statement or for earnings releases, corporate actions, and so on?
“More tagged data makes it cheaper and easier for investors and investment professionals to compare across companies.”
Every time there is a new form, or every time a form is revised, DERA (the Office of Structured Disclosure in particular) will look at whether there is a good role for tagging or structuring the data that is being requested. The proposed pay-versus-performance rules, if they are approved by the Commission, will be the first use of XBRL tagging in the proxy statement. I believe that more tagged data generally is better. More tagged data makes it cheaper and easier for investors and investment professionals to compare across companies. We would like to have more of that wherever possible, after considering the corresponding benefits and costs. DERA is certainly a big fan of increasing the amount of tagging to the extent that the Commission thinks is appropriate.
What plans does the SEC have to promote the use of XBRL data by investors and others in the market?
We do not have plans to promote XBRL per se, but we have plans to promote what we think is the most appropriate version of tagged data. We have, as I said, posted about 19 or 20 quarters of face financial data for SEC-registered firms. Because they are tagged in XBRL, we are able to get that data out quickly. We hope and expect that the public will understand that one of the reasons why they get the data quickly is the tagging. As we collect more information, as there are new or revised rules, we will always consider structuring the related data.
How would you respond to a CFO who wonders whether XBRL tagging is worth the necessary time and expense?
We always try to take into consideration the fact that our reporting obligations entail costs, and we try very hard not to impose costs that we do not think are justifiable. We hear about costs from small firms in some cases because many of the reporting expenses are fixed costs that are independent of company size. The big advantage of having tagged data for small companies is that - because it is cheaper for analysts to use that data - more analysts will use the data; and the more people in the markets who know about a small company, the deeper the market for its stock will be. The presence of readily available data about small companies, those outside the traditional sources of data historically provided by the commercial aggregators, will make a difference in the demand for their stocks. Outsiders will be able to understand a small company better, put it in perspective, understand the extent to which it adds risk or diversification to their portfolios. The other important point is to emphasize the rapid increase in the number of service providers in this area. The cost has been falling, the benefit continues to be pretty high in the tradability of the stock, and, over time, XBRL will become less of an issue in the CFO's concern for the bottom line.
NOTE: The views expressed here are entirely Dr. Flannery's, and they do not necessarily reflect those of the SEC, members of the SEC staff, or any other organization.
Click here to access all Dimensions eNewsletters